fix(handler): make on_comment and @mention paths mutually exclusive

Without (issue, agent) coalescing, a single member comment that @mentions
the assignee was enqueueing two tasks with identical trigger_comment_id —
once via shouldEnqueueOnComment → EnqueueTaskForIssue and once via
enqueueMentionedAgentTasks → EnqueueTaskForMention. The same double-fire
hit plain replies that inherit the assignee mention from the thread root.

Add commentMentionsAssignee, which uses the shared
shouldInheritParentMentions logic to compute the same effective mention
set the @mention path will see. Extend the on_comment gate to skip when
that helper says the assignee will be triggered through the @mention
path. Net result: exactly one task per (comment, assignee), with the
trigger comment preserved.

TestOnCommentTriggerDecision updated to reflect the new contract; assignee
mention cases that used to assert the on_comment branch fires now assert
it skips (mention path covers them). New integration test
TestAssigneeMentionDoesNotDoubleEnqueue pins the end-to-end behavior.

Also softens the "ClaimAgentTask guarantees serial execution" wording in
the code and the RFC: it is a coordination-side property under the
current single-poller-per-runtime model, not a DB-level lock on
(issue, agent). FOR UPDATE SKIP LOCKED locks the row being claimed, not
the key.

Co-authored-by: multica-agent <github@multica.ai>

.env.example

@@ -63,6 +63,9 @@ GOOGLE_CLIENT_SECRET=
 GOOGLE_REDIRECT_URI=http://localhost:3000/auth/callback
 
 # S3 / CloudFront
+# S3_BUCKET — bucket NAME only (e.g. "my-bucket"). Do NOT include the
+# ".s3.<region>.amazonaws.com" suffix; the server builds the public URL
+# from S3_BUCKET + S3_REGION. S3_REGION must match the bucket's real region.
 S3_BUCKET=
 S3_REGION=us-west-2
 CLOUDFRONT_KEY_PAIR_ID=
@@ -129,5 +132,8 @@ ALLOWED_EMAILS=
 # will run a no-op analytics client and ship nothing. See docs/analytics.md.
 POSTHOG_API_KEY=
 POSTHOG_HOST=https://us.i.posthog.com
+# Optional override for the `environment` PostHog event property.
+# Defaults from APP_ENV and normalizes to production / staging / dev.
+ANALYTICS_ENVIRONMENT=
 # Force the no-op client even when POSTHOG_API_KEY is set (CI / opt-out).
 ANALYTICS_DISABLED=

.github/PULL_REQUEST_TEMPLATE.md

@@ -40,6 +40,8 @@ Closes #
 - [ ] I have added or updated tests where applicable
 - [ ] If this change affects the UI, I have included before/after screenshots
 - [ ] I have updated relevant documentation to reflect my changes
+- [ ] If I added a new runtime / coding tool / UI tab, I synced the change to **landing copy** (`apps/web/features/landing/i18n/`), **starter-content** (`packages/views/onboarding/utils/starter-content-content-*.ts`), and **relevant docs** (`apps/docs/content/docs/`)
+- [ ] If this PR touches Chinese product copy, I checked it against `apps/docs/content/docs/developers/conventions.zh.mdx` (terminology, mixed-rule for `task` / `issue` / `skill`)
 - [ ] I have considered and documented any risks above
 - [ ] I will address all reviewer comments before requesting merge
 

.github/workflows/ci.yml

@@ -29,8 +29,17 @@ jobs:
       - name: Install dependencies
         run: pnpm install
 
-      - name: Build, type check, and test
-        run: pnpm exec turbo build typecheck test --filter='!@multica/docs'
+      - name: Verify reserved-slugs.ts is up to date
+        # Re-runs the generator and fails on any drift from the
+        # checked-in TypeScript output. The Go side embeds the JSON
+        # source directly, so a passing diff here proves both sides
+        # share one source of truth.
+        run: |
+          pnpm generate:reserved-slugs
+          git diff --exit-code -- packages/core/paths/reserved-slugs.ts
+
+      - name: Build, type check, lint, and test
+        run: pnpm exec turbo build typecheck lint test --filter='!@multica/docs'
 
   backend:
     runs-on: ubuntu-latest

CLAUDE.md

@@ -2,6 +2,21 @@
 
 This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
 
+## Conventions reference
+
+The single source of truth for **code naming, the i18n translation glossary, and the Chinese voice guide** is the docs site:
+
+- **`apps/docs/content/docs/developers/conventions.mdx`** (English)
+- **`apps/docs/content/docs/developers/conventions.zh.mdx`** (Chinese)
+
+Read that page before:
+
+- Writing or editing translations (`packages/views/locales/`)
+- Naming a new route, package, file, DB column, or TS type
+- Writing Chinese product copy (UI strings, error messages, docs)
+
+The legacy `packages/views/locales/glossary.md` is now a stub redirecting to the docs page; do not rely on it.
+
 ## Project Context
 
 Multica is an AI-native task management platform — like Linear, but with AI agents as first-class citizens.
@@ -131,10 +146,27 @@ make start-worktree     # Start using .env.worktree
 - Go code follows standard Go conventions (gofmt, go vet).
 - Keep comments in code **English only**.
 - Prefer existing patterns/components over introducing parallel abstractions.
-- Unless the user explicitly asks for backwards compatibility, do **not** add compatibility layers, fallback paths, dual-write logic, legacy adapters, or temporary shims.
+- Unless the user explicitly asks for backwards compatibility, do **not** add compatibility layers, fallback paths, dual-write logic, legacy adapters, or temporary shims **for internal, non-boundary code** (a function calling another function in the same package, a component reading its own state, a store helper, etc.).
+- This rule does **not** apply at API boundaries: the desktop app cannot assume the backend it talks to has the same shape as the one it was built against (older desktop installs will outlive any given server build). API response handling must follow the rules in **API Response Compatibility** below — that is a defensive boundary, not a legacy shim.
 - If a flow or API is being replaced and the product is not yet live, prefer removing the old path instead of preserving both old and new behavior.
 - Avoid broad refactors unless required by the task.
 - New global (pre-workspace) routes MUST use a single word (`/login`, `/inbox`) or a `/{noun}/{verb}` pair (`/workspaces/new`). NEVER add hyphenated word-group root routes (`/new-workspace`, `/create-team`) — they collide with common user workspace names and force endless reserved-slug audits. Reserving the noun (`workspaces`) automatically protects the entire `/workspaces/*` subtree.
+- The reserved-slug list lives in **one** place: `server/internal/handler/reserved_slugs.json`. The Go side embeds the JSON; `packages/core/paths/reserved-slugs.ts` is generated from it by `pnpm generate:reserved-slugs`. Edit the JSON, run the generator, commit both. CI re-runs the generator and fails on any drift, so a stale TS file cannot land.
+
+### API Response Compatibility
+
+The desktop app installed on a user's machine is older than any backend it talks to: a user on 0.2.26 will hit a server running 0.3.x, then 0.4.x, then beyond. Every response shape is a contract that **will** drift, and the frontend must survive drift without white-screening. Three concrete incidents already happened from violating this — #2143, #2147, #2192.
+
+When writing code that consumes an API response, follow these rules:
+
+- **Parse, don't cast.** Untyped JSON crossing the network is not `T`. Use `parseWithFallback` in `packages/core/api/schema.ts` with a `zod` schema and an explicit fallback. On validation failure it logs a warning and returns the fallback; it never throws into the UI.
+- **No bare `as` casts on response bodies.** Every endpoint method whose response is consumed by UI logic must run through a schema before returning.
+- **Optional-chain and default everywhere downstream.** Treat every field as possibly missing. Use explicit boolean checks (`=== true`) over truthy/falsy negation, which silently treats `undefined` and `null` as `false`.
+- **Don't pin a UI affordance to a single backend field.** If a button or indicator depends on exactly one boolean from the server, a backend bug deletes it. Combine signals (cursor presence, page length, etc.) so the affordance stays available in the worst case.
+- **Enum drift downgrades, not crashes.** A new server-side enum value should render a generic fallback. `switch` statements on server-driven strings must have a `default` branch.
+- **When you add or change an endpoint:** add the schema in the same PR, and write at least one test that feeds a malformed response through it (missing field, wrong type, `null` array). The test fails closed if a future change breaks the contract.
+
+This is not premature defense — it is the *only* defense for an installed-app architecture. CSR-only browser apps can ship a fix in minutes; an Electron build sitting on a developer's laptop cannot.
 
 ### Backend Handler UUID Parsing Convention
 

CLI_AND_DAEMON.md

@@ -70,10 +70,10 @@ Opens your browser for OAuth authentication, creates a 90-day personal access to
 ### Token Login
 
 ```bash
-multica login --token
+multica login --token <mul_...>
 ```
 
-Authenticate by pasting a personal access token directly. Useful for headless environments.
+Authenticate using a personal access token directly. Useful for headless environments. Pass `--token=` with an empty value to be prompted interactively (so the token never lands in shell history).
 
 ### Check Status
 
@@ -140,6 +140,7 @@ The daemon auto-detects these AI CLIs on your PATH:
 |-----|---------|-------------|
 | [Claude Code](https://docs.anthropic.com/en/docs/claude-code) | `claude` | Anthropic's coding agent |
 | [Codex](https://github.com/openai/codex) | `codex` | OpenAI's coding agent |
+| [GitHub Copilot CLI](https://docs.github.com/en/copilot) | `copilot` | GitHub's coding agent (model routed by your GitHub entitlement) |
 | OpenCode | `opencode` | Open-source coding agent |
 | OpenClaw | `openclaw` | Open-source coding agent |
 | Hermes | `hermes` | Nous Research coding agent |
@@ -174,6 +175,22 @@ Daemon behavior is configured via flags or environment variables:
 | Device name | `--device-name` | `MULTICA_DAEMON_DEVICE_NAME` | hostname |
 | Runtime name | `--runtime-name` | `MULTICA_AGENT_RUNTIME_NAME` | `Local Agent` |
 | Workspaces root | — | `MULTICA_WORKSPACES_ROOT` | `~/multica_workspaces` |
+| GC enabled | — | `MULTICA_GC_ENABLED` | `true` (set `false`/`0` to disable) |
+| GC scan interval | — | `MULTICA_GC_INTERVAL` | `1h` |
+| GC TTL (done/cancelled issues) | — | `MULTICA_GC_TTL` | `24h` |
+| GC orphan TTL (no `.gc_meta.json`) | — | `MULTICA_GC_ORPHAN_TTL` | `72h` |
+| GC artifact TTL (open issues) | — | `MULTICA_GC_ARTIFACT_TTL` | `12h` (set `0` to disable) |
+| GC artifact patterns | — | `MULTICA_GC_ARTIFACT_PATTERNS` | `node_modules,.next,.turbo` |
+
+#### Workspace garbage collection
+
+The daemon periodically scans `MULTICA_WORKSPACES_ROOT` and reclaims disk space in three modes:
+
+- **Full task cleanup** — when an issue's status is `done` or `cancelled` and has been idle for `MULTICA_GC_TTL`, the entire task directory is removed.
+- **Orphan cleanup** — task directories with no `.gc_meta.json` (e.g. left over from a daemon crash) are removed once they exceed `MULTICA_GC_ORPHAN_TTL`.
+- **Artifact-only cleanup** — when a task has been completed for at least `MULTICA_GC_ARTIFACT_TTL` but the issue is still open, regenerable build outputs whose directory basename matches `MULTICA_GC_ARTIFACT_PATTERNS` are removed; the rest of the workdir (source, `.git`, `output/`, `logs/`, `.gc_meta.json`) is preserved so the agent can resume the same workdir on the next task.
+
+Patterns are basename-only — entries containing `/` or `\` are silently dropped — and `.git` subtrees are never descended into. The default list (`node_modules`, `.next`, `.turbo`) is intentionally narrow; extend it per deployment if your repos consistently produce other regenerable directories (for example, `MULTICA_GC_ARTIFACT_PATTERNS=node_modules,.next,.turbo,target,__pycache__`). To disable artifact cleanup entirely, set `MULTICA_GC_ARTIFACT_TTL=0`.
 
 Agent-specific overrides:
 
@@ -181,8 +198,12 @@ Agent-specific overrides:
 |----------|-------------|
 | `MULTICA_CLAUDE_PATH` | Custom path to the `claude` binary |
 | `MULTICA_CLAUDE_MODEL` | Override the Claude model used |
+| `MULTICA_CLAUDE_ARGS` | Default extra arguments for Claude Code runs |
 | `MULTICA_CODEX_PATH` | Custom path to the `codex` binary |
 | `MULTICA_CODEX_MODEL` | Override the Codex model used |
+| `MULTICA_CODEX_ARGS` | Default extra arguments for Codex runs |
+| `MULTICA_COPILOT_PATH` | Custom path to the `copilot` binary |
+| `MULTICA_COPILOT_MODEL` | Override the Copilot model used (note: GitHub Copilot routes models through your account entitlement, so this may not be honoured) |
 | `MULTICA_OPENCODE_PATH` | Custom path to the `opencode` binary |
 | `MULTICA_OPENCODE_MODEL` | Override the OpenCode model used |
 | `MULTICA_OPENCLAW_PATH` | Custom path to the `openclaw` binary |
@@ -200,6 +221,8 @@ Agent-specific overrides:
 | `MULTICA_KIRO_PATH` | Custom path to the `kiro-cli` binary |
 | `MULTICA_KIRO_MODEL` | Override the Kiro model used |
 
+`MULTICA_CLAUDE_ARGS` and `MULTICA_CODEX_ARGS` are parsed with POSIX shellword quoting, so values such as `--model "gpt-5.1 codex" --sandbox read-only` are split like a shell command line. Agent arguments are applied in this order: hardcoded Multica defaults, daemon-wide env defaults, then per-agent `custom_args` from the task.
+
 ### Self-Hosted Server
 
 When connecting to a self-hosted Multica instance, the easiest approach is:
@@ -282,10 +305,12 @@ multica workspace members <workspace-id>
 multica issue list
 multica issue list --status in_progress
 multica issue list --priority urgent --assignee "Agent Name"
+multica issue list --assignee-id 5fb87ac7-23b5-4a7a-81fa-ed295a54545d
+multica issue list --full-id
 multica issue list --limit 20 --output json
 ```
 
-Available filters: `--status`, `--priority`, `--assignee`, `--project`, `--limit`.
+Table output shows a routable issue `KEY` such as `MUL-123`; copy that key into follow-up commands like `issue get`, `issue comment list`, `issue status`, or `--parent`. Add `--full-id` when you need canonical UUIDs. Available filters: `--status`, `--priority`, `--assignee` / `--assignee-id`, `--project`, `--limit`. Use `--assignee-id <uuid>` for unambiguous filtering when names overlap.
 
 ### Get Issue
 
@@ -298,9 +323,10 @@ multica issue get <id> --output json
 
 ```bash
 multica issue create --title "Fix login bug" --description "..." --priority high --assignee "Lambda"
+multica issue create --title "Fix login bug" --assignee-id 5fb87ac7-23b5-4a7a-81fa-ed295a54545d
 ```
 
-Flags: `--title` (required), `--description`, `--status`, `--priority`, `--assignee`, `--parent`, `--project`, `--due-date`.
+Flags: `--title` (required), `--description`, `--status`, `--priority`, `--assignee` / `--assignee-id`, `--parent`, `--project`, `--due-date`. Pass `--assignee-id <uuid>` (mutually exclusive with `--assignee`) when scripting against the IDs returned by `multica workspace members --output json` / `multica agent list --output json`.
 
 ### Update Issue
 
@@ -312,9 +338,12 @@ multica issue update <id> --title "New title" --priority urgent
 
 ```bash
 multica issue assign <id> --to "Lambda"
+multica issue assign <id> --to-id 5fb87ac7-23b5-4a7a-81fa-ed295a54545d
 multica issue assign <id> --unassign
 ```
 
+Pass `--to-id <uuid>` to assign by canonical UUID (mutually exclusive with `--to`); useful when names overlap across members and agents.
+
 ### Change Status
 
 ```bash
@@ -365,17 +394,19 @@ Subscribers receive notifications about issue activity (new comments, status cha
 ```bash
 # List all execution runs for an issue
 multica issue runs <issue-id>
+multica issue runs <issue-id> --full-id
 multica issue runs <issue-id> --output json
 
 # View messages for a specific execution run
 multica issue run-messages <task-id>
+multica issue run-messages <short-task-id> --issue <issue-id>
 multica issue run-messages <task-id> --output json
 
 # Incremental fetch (only messages after a given sequence number)
 multica issue run-messages <task-id> --since 42 --output json
 ```
 
-The `runs` command shows all past and current executions for an issue, including running tasks. The `run-messages` command shows the detailed message log (tool calls, thinking, text, errors) for a single run. Use `--since` for efficient polling of in-progress runs.
+The `runs` command shows all past and current executions for an issue, including running tasks. Table output uses short task UUID prefixes by default; pass `--full-id` to print canonical task UUIDs. The `run-messages` command accepts full task UUIDs directly; copied short task prefixes must be scoped with `--issue <issue-id>` so the CLI only checks that issue's runs. It shows the detailed message log (tool calls, thinking, text, errors) for a single run. Use `--since` for efficient polling of in-progress runs.
 
 ## Projects
 
@@ -485,9 +516,12 @@ Autopilots are scheduled/triggered automations that dispatch agent tasks (either
 
 ```bash
 multica autopilot list
+multica autopilot list --full-id
 multica autopilot list --status active --output json
 ```
 
+Autopilot table IDs are short UUID prefixes; follow-up autopilot commands accept copied prefixes when they are unique in the current workspace. Use `--full-id` to print canonical UUIDs.
+
 ### Get Autopilot Details
 
 ```bash

CLI_INSTALL.md

@@ -140,7 +140,7 @@ multica auth status
 Expected output should show the authenticated user and server URL.
 
 **If login fails:**
-- If no browser is available (headless environment), the user can generate a Personal Access Token at `https://app.multica.ai/settings` and run: `multica login --token`
+- If no browser is available (headless environment), the user can generate a Personal Access Token at `https://app.multica.ai/settings` and run: `multica login --token <mul_...>` (use `--token=` with an empty value to be prompted interactively).
 - If the server URL needs to be customized: `multica config set server_url <url>` before logging in.
 
 ---
@@ -166,12 +166,12 @@ Wait 3 seconds, then verify:
 multica daemon status
 ```
 
-Expected output should show `running` status with detected agents (e.g. `claude`, `codex`, `opencode`, `openclaw`, `hermes`, `gemini`, `pi`, `cursor-agent`).
+Expected output should show `running` status with detected agents (e.g. `claude`, `codex`, `copilot`, `opencode`, `openclaw`, `hermes`, `gemini`, `pi`, `cursor-agent`).
 
 **If daemon fails to start:**
 - Check logs: `multica daemon logs`
 - If a port conflict occurs, the daemon may already be running under a different profile.
-- If no agents are detected, ensure at least one AI CLI (`claude`, `codex`, `opencode`, `openclaw`, `hermes`, `gemini`, `pi`, or `cursor-agent`) is installed and on the `$PATH`.
+- If no agents are detected, ensure at least one AI CLI (`claude`, `codex`, `copilot`, `opencode`, `openclaw`, `hermes`, `gemini`, `pi`, or `cursor-agent`) is installed and on the `$PATH`.
 
 ---
 
@@ -185,12 +185,12 @@ multica daemon status
 
 Confirm:
 1. Status is `running`
-2. At least one agent is listed (e.g. `claude`, `codex`, `opencode`, `openclaw`, `hermes`, `gemini`, `pi`, or `cursor-agent`)
+2. At least one agent is listed (e.g. `claude`, `codex`, `copilot`, `opencode`, `openclaw`, `hermes`, `gemini`, `pi`, or `cursor-agent`)
 3. At least one workspace is being watched
 
 If the agents list is empty, tell the user:
 
-> "The Multica daemon is running but no AI agent CLIs were detected. Please install at least one supported CLI (`claude`, `codex`, `opencode`, `openclaw`, `hermes`, `gemini`, `pi`, or `cursor-agent`), then restart the daemon with `multica daemon stop && multica daemon start`."
+> "The Multica daemon is running but no AI agent CLIs were detected. Please install at least one supported CLI (`claude`, `codex`, `copilot`, `opencode`, `openclaw`, `hermes`, `gemini`, `pi`, or `cursor-agent`), then restart the daemon with `multica daemon stop && multica daemon start`."
 
 ---
 

README.md

@@ -30,12 +30,24 @@ Turn coding agents into real teammates — assign tasks, track progress, compoun
 
 Multica turns coding agents into real teammates. Assign issues to an agent like you'd assign to a colleague — they'll pick up the work, write code, report blockers, and update statuses autonomously.
 
-No more copy-pasting prompts. No more babysitting runs. Your agents show up on the board, participate in conversations, and compound reusable skills over time. Think of it as open-source infrastructure for managed agents — vendor-neutral, self-hosted, and designed for human + AI teams. Works with **Claude Code**, **Codex**, **OpenClaw**, **OpenCode**, **Hermes**, **Gemini**, **Pi**, **Cursor Agent**, **Kimi**, and **Kiro CLI**.
+No more copy-pasting prompts. No more babysitting runs. Your agents show up on the board, participate in conversations, and compound reusable skills over time. Think of it as open-source infrastructure for managed agents — vendor-neutral, self-hosted, and designed for human + AI teams. Works with **Claude Code**, **Codex**, **GitHub Copilot CLI**, **OpenClaw**, **OpenCode**, **Hermes**, **Gemini**, **Pi**, **Cursor Agent**, **Kimi**, and **Kiro CLI**.
 
 <p align="center">
   <img src="docs/assets/hero-screenshot.png" alt="Multica board view" width="800">
 </p>
 
+## Why "Multica"?
+
+Multica — **Mul**tiplexed **I**nformation and **C**omputing **A**gent.
+
+The name is a nod to Multics, the pioneering operating system of the 1960s that introduced time-sharing — letting multiple users share a single machine as if each had it to themselves. Unix was born as a deliberate simplification of Multics: one user, one task, one elegant philosophy.
+
+We think the same inflection is happening again. For decades, software teams have been single-threaded — one engineer, one task, one context switch at a time. AI agents change that equation. Multica brings time-sharing back, but for an era where the "users" multiplexing the system are both humans and autonomous agents.
+
+In Multica, agents are first-class teammates. They get assigned issues, report progress, raise blockers, and ship code — just like their human colleagues. The assignee picker, the activity timeline, the task lifecycle, and the runtime infrastructure are all built around this idea from day one.
+
+Like Multics before it, the bet is on multiplexing: a small team shouldn't feel small. With the right system, two engineers and a fleet of agents can move like twenty.
+
 ## Features
 
 Multica manages the full agent lifecycle: from task assignment to execution monitoring to skill reuse.
@@ -98,7 +110,7 @@ multica setup          # Connect to Multica Cloud, log in, start daemon
 multica setup           # Configure, authenticate, and start the daemon
 ```
 
-The daemon runs in the background and auto-detects agent CLIs (`claude`, `codex`, `openclaw`, `opencode`, `hermes`, `gemini`, `pi`, `cursor-agent`, `kimi`, `kiro-cli`) on your PATH.
+The daemon runs in the background and auto-detects agent CLIs (`claude`, `codex`, `copilot`, `openclaw`, `opencode`, `hermes`, `gemini`, `pi`, `cursor-agent`, `kimi`, `kiro-cli`) on your PATH.
 
 ### 2. Verify your runtime
 
@@ -108,7 +120,7 @@ Open your workspace in the Multica web app. Navigate to **Settings → Runtimes*
 
 ### 3. Create an agent
 
-Go to **Settings → Agents** and click **New Agent**. Pick the runtime you just connected and choose a provider (Claude Code, Codex, OpenClaw, OpenCode, Hermes, Gemini, Pi, Cursor Agent, Kimi, or Kiro CLI). Give your agent a name — this is how it will appear on the board, in comments, and in assignments.
+Go to **Settings → Agents** and click **New Agent**. Pick the runtime you just connected and choose a provider (Claude Code, Codex, GitHub Copilot CLI, OpenClaw, OpenCode, Hermes, Gemini, Pi, Cursor Agent, Kimi, or Kiro CLI). Give your agent a name — this is how it will appear on the board, in comments, and in assignments.
 
 ### 4. Assign your first task
 
@@ -160,10 +172,9 @@ See the [CLI and Daemon Guide](CLI_AND_DAEMON.md) for the full command reference
                             │
                      ┌──────┴───────┐
                      │ Agent Daemon │  runs on your machine
-                     └──────────────┘  (Claude Code, Codex, OpenCode,
-                                        OpenClaw, Hermes, Gemini,
-                                        Pi, Cursor Agent, Kimi,
-                                        Kiro CLI)
+                     └──────────────┘  (Claude Code, Codex, GitHub Copilot CLI,
+                                        OpenCode, OpenClaw, Hermes, Gemini,
+                                        Pi, Cursor Agent, Kimi, Kiro CLI)
 ```
 
 | Layer | Stack |
@@ -171,7 +182,7 @@ See the [CLI and Daemon Guide](CLI_AND_DAEMON.md) for the full command reference
 | Frontend | Next.js 16 (App Router) |
 | Backend | Go (Chi router, sqlc, gorilla/websocket) |
 | Database | PostgreSQL 17 with pgvector |
-| Agent Runtime | Local daemon executing Claude Code, Codex, OpenClaw, OpenCode, Hermes, Gemini, Pi, Cursor Agent, Kimi, or Kiro CLI |
+| Agent Runtime | Local daemon executing Claude Code, Codex, GitHub Copilot CLI, OpenClaw, OpenCode, Hermes, Gemini, Pi, Cursor Agent, Kimi, or Kiro CLI |
 
 ## Development
 

README.zh-CN.md

@@ -30,12 +30,24 @@
 
 Multica 将编码 Agent 变成真正的队友。像分配给同事一样分配给 Agent——它们会自主接手工作、编写代码、报告阻塞问题、更新状态。
 
-不再需要复制粘贴 prompt，不再需要盯着运行过程。你的 Agent 出现在看板上、参与对话、随着时间积累可复用的技能。可以理解为开源的 Managed Agents 基础设施——厂商中立、可自部署、专为人类 + AI 团队设计。支持 **Claude Code**、**Codex**、**OpenClaw**、**OpenCode**、**Hermes**、**Gemini**、**Pi** 和 **Cursor Agent**。
+不再需要复制粘贴 prompt，不再需要盯着运行过程。你的 Agent 出现在看板上、参与对话、随着时间积累可复用的技能。可以理解为开源的 Managed Agents 基础设施——厂商中立、可自部署、专为人类 + AI 团队设计。支持 **Claude Code**、**Codex**、**GitHub Copilot CLI**、**OpenClaw**、**OpenCode**、**Hermes**、**Gemini**、**Pi**、**Cursor Agent**、**Kimi** 和 **Kiro CLI**。
 
 <p align="center">
   <img src="docs/assets/hero-screenshot.png" alt="Multica 看板视图" width="800">
 </p>
 
+## 为什么叫 "Multica"？
+
+Multica——**Mul**tiplexed **I**nformation and **C**omputing **A**gent。
+
+这个名字是在向 20 世纪 60 年代具有开创意义的操作系统 Multics 致意。Multics 首创了分时系统，让多个用户能够共享同一台机器，同时又像各自独占它一样使用。Unix 则是在有意简化 Multics 的基础上诞生的，强调一个用户、一个任务、一种优雅的哲学。
+
+我们认为，类似的转折点正在再次出现。几十年来，软件团队一直处于一种单线程的工作模式，一个工程师处理一个任务，一次只专注于一个上下文。AI agents 改变了这个等式。Multica 将"分时"重新带回这个时代，只不过今天在系统中进行多路复用的"用户"，既包括人类，也包括自主代理。
+
+在 Multica 中，agents 是一级团队成员。它们会被分配 issue，汇报进展，提出阻塞，并交付代码，就像人类同事一样。任务分配、活动时间线、任务生命周期，以及运行时基础设施，Multica 从第一天起就是围绕这一理念构建的。
+
+和当年的 Multics 一样，这一判断建立在"多路复用"之上。一个小团队不该因为人数少就显得能力有限。有了合适的系统，两名工程师加上一组 agents，就能发挥出二十人团队的推进速度。
+
 ## 功能特性
 
 Multica 管理完整的 Agent 生命周期：从任务分配到执行监控再到技能复用。
@@ -99,7 +111,7 @@ multica setup          # 连接 Multica Cloud，登录，启动 daemon
 multica setup           # 配置、认证、启动 daemon（一条命令搞定）
 ```
 
-daemon 在后台运行，保持你的机器与 Multica 的连接。它会自动检测 PATH 中可用的 Agent CLI（`claude`、`codex`、`openclaw`、`opencode`、`hermes`、`gemini`、`pi`、`cursor-agent`）。
+daemon 在后台运行，保持你的机器与 Multica 的连接。它会自动检测 PATH 中可用的 Agent CLI（`claude`、`codex`、`copilot`、`openclaw`、`opencode`、`hermes`、`gemini`、`pi`、`cursor-agent`、`kimi`、`kiro-cli`）。
 
 ### 2. 确认运行时已连接
 
@@ -109,7 +121,7 @@ daemon 在后台运行，保持你的机器与 Multica 的连接。它会自动
 
 ### 3. 创建 Agent
 
-进入 **设置 → Agents**，点击 **新建 Agent**。选择你刚连接的 Runtime，选择 Provider（Claude Code、Codex、OpenClaw、OpenCode、Hermes、Gemini、Pi 或 Cursor Agent），并为 Agent 起个名字——它将以这个名字出现在看板、评论和任务分配中。
+进入 **设置 → Agents**，点击 **新建 Agent**。选择你刚连接的 Runtime，选择 Provider（Claude Code、Codex、GitHub Copilot CLI、OpenClaw、OpenCode、Hermes、Gemini、Pi、Cursor Agent、Kimi 或 Kiro CLI），并为 Agent 起个名字——它将以这个名字出现在看板、评论和任务分配中。
 
 ### 4. 分配你的第一个任务
 
@@ -142,9 +154,9 @@ daemon 在后台运行，保持你的机器与 Multica 的连接。它会自动
                             │
                      ┌──────┴───────┐
                      │ Agent Daemon │  运行在你的机器上
-                     └──────────────┘  （Claude Code、Codex、OpenCode、
-                                        OpenClaw、Hermes、Gemini、
-                                        Pi、Cursor Agent）
+                     └──────────────┘  （Claude Code、Codex、GitHub Copilot CLI、
+                                        OpenCode、OpenClaw、Hermes、Gemini、
+                                        Pi、Cursor Agent、Kimi、Kiro CLI）
 ```
 
 | 层级 | 技术栈 |
@@ -152,7 +164,7 @@ daemon 在后台运行，保持你的机器与 Multica 的连接。它会自动
 | 前端 | Next.js 16 (App Router) |
 | 后端 | Go (Chi router, sqlc, gorilla/websocket) |
 | 数据库 | PostgreSQL 17 with pgvector |
-| Agent 运行时 | 本地 daemon 执行 Claude Code、Codex、OpenClaw、OpenCode、Hermes、Gemini、Pi 或 Cursor Agent |
+| Agent 运行时 | 本地 daemon 执行 Claude Code、Codex、GitHub Copilot CLI、OpenClaw、OpenCode、Hermes、Gemini、Pi、Cursor Agent、Kimi 或 Kiro CLI |
 
 ## 开发
 

SELF_HOSTING.md

@@ -92,6 +92,7 @@ brew install multica-ai/tap/multica
 You also need at least one AI agent CLI installed:
 - [Claude Code](https://docs.anthropic.com/en/docs/claude-code) (`claude` on PATH)
 - [Codex](https://github.com/openai/codex) (`codex` on PATH)
+- [GitHub Copilot CLI](https://docs.github.com/en/copilot) (`copilot` on PATH)
 - [OpenClaw](https://github.com/openclaw/openclaw) (`openclaw` on PATH)
 - [OpenCode](https://github.com/anomalyco/opencode) (`opencode` on PATH)
 - [Hermes](https://github.com/NousResearch/hermes) (`hermes` on PATH)

SELF_HOSTING_ADVANCED.md

@@ -56,13 +56,15 @@ Changes take effect after restarting the backend / compose stack. The web UI rea
 
 ### File Storage (Optional)
 
-For file uploads and attachments, configure S3 and CloudFront:
+For file uploads and attachments, configure S3 and (optionally) CloudFront:
 
 | Variable | Description |
 |----------|-------------|
-| `S3_BUCKET` | S3 bucket name |
-| `S3_REGION` | AWS region (default: `us-west-2`) |
-| `CLOUDFRONT_DOMAIN` | CloudFront distribution domain |
+| `S3_BUCKET` | Bucket name only (e.g. `my-bucket`). Do **not** include the `.s3.<region>.amazonaws.com` suffix — the server constructs the public URL from `S3_BUCKET` + `S3_REGION` |
+| `S3_REGION` | AWS region (default: `us-west-2`). Must match the bucket's actual region — used for both SDK signing and public URLs |
+| `AWS_ACCESS_KEY_ID` / `AWS_SECRET_ACCESS_KEY` | Static credentials. When both are unset, the AWS SDK default credential chain is used |
+| `AWS_ENDPOINT_URL` | Custom S3-compatible endpoint (e.g. MinIO, R2, B2). Setting this switches the public URL to path-style |
+| `CLOUDFRONT_DOMAIN` | CloudFront distribution domain — when set, public URLs use this host instead of the S3 host |
 | `CLOUDFRONT_KEY_PAIR_ID` | CloudFront key pair ID for signed URLs |
 | `CLOUDFRONT_PRIVATE_KEY` | CloudFront private key (PEM format) |
 
@@ -103,6 +105,8 @@ Agent-specific overrides:
 | `MULTICA_CLAUDE_MODEL` | Override the Claude model used |
 | `MULTICA_CODEX_PATH` | Custom path to the `codex` binary |
 | `MULTICA_CODEX_MODEL` | Override the Codex model used |
+| `MULTICA_COPILOT_PATH` | Custom path to the `copilot` (GitHub Copilot CLI) binary |
+| `MULTICA_COPILOT_MODEL` | Override the Copilot model used (note: GitHub Copilot routes models through your account entitlement, so this may not be honoured) |
 | `MULTICA_OPENCODE_PATH` | Custom path to the `opencode` binary |
 | `MULTICA_OPENCODE_MODEL` | Override the OpenCode model used |
 | `MULTICA_OPENCLAW_PATH` | Custom path to the `openclaw` binary |

apps/desktop/.env.production

@@ -1,12 +0,0 @@
-# Production environment for `pnpm package` / `pnpm build`.
-# electron-vite (Vite under the hood) reads this automatically in
-# production mode and inlines the values into the renderer bundle via
-# import.meta.env.VITE_*. These are public URLs, not secrets.
-
-# Backend API + websocket the desktop app talks to.
-VITE_API_URL=https://api.multica.ai
-VITE_WS_URL=wss://api.multica.ai/ws
-
-# Public web app URL — used to build shareable links like "Copy link to
-# issue" that users paste into Slack / messages. See platform/navigation.tsx.
-VITE_APP_URL=https://multica.ai

apps/desktop/src/main/app-version.ts

@@ -0,0 +1,33 @@
+import { app } from "electron";
+import { execSync } from "node:child_process";
+
+/**
+ * Resolve the running app version. In packaged builds this is the value
+ * `electron-builder` baked into package.json via `extraMetadata.version`
+ * (driven by `git describe` — see `apps/desktop/scripts/package.mjs`), so
+ * `app.getVersion()` matches the GitHub Release tag exactly.
+ *
+ * In dev (`pnpm dev:desktop`) `app.getVersion()` only sees the static
+ * `apps/desktop/package.json` value, which is "0.1.0" and never bumped —
+ * the Settings → Updates panel and any other UI surfacing the version
+ * would mislead developers into thinking they're running ancient builds.
+ * Fall back to `git describe --tags --always --dirty` (same source the
+ * packager uses) so dev shows e.g. `0.2.19-14-gabcdef-dirty`. If git is
+ * unavailable for whatever reason, we just return the package.json value.
+ */
+export function getAppVersion(): string {
+  if (app.isPackaged) {
+    return app.getVersion();
+  }
+  try {
+    const raw = execSync("git describe --tags --always --dirty", {
+      cwd: app.getAppPath(),
+      encoding: "utf-8",
+      stdio: ["ignore", "pipe", "ignore"],
+    }).trim();
+    if (!raw) return app.getVersion();
+    return raw.replace(/^v/, "");
+  } catch {
+    return app.getVersion();
+  }
+}

apps/desktop/src/main/index.ts

@@ -7,6 +7,9 @@ import { setupAutoUpdater } from "./updater";
 import { setupDaemonManager } from "./daemon-manager";
 import { openExternalSafely } from "./external-url";
 import { installContextMenu } from "./context-menu";
+import { getAppVersion } from "./app-version";
+import { loadRuntimeConfig } from "./runtime-config-loader";
+import type { RuntimeConfigResult } from "../shared/runtime-config";
 
 // Bundled icon used for dev-mode dock/taskbar branding. In production the
 // app bundle icon (from electron-builder) wins; this path is only consumed
@@ -36,6 +39,10 @@ if (process.platform !== "win32") {
 const PROTOCOL = "multica";
 
 let mainWindow: BrowserWindow | null = null;
+let runtimeConfigResult: RuntimeConfigResult = {
+  ok: false,
+  error: { message: "Runtime config has not loaded yet" },
+};
 
 // --- Deep link helpers ---------------------------------------------------
 
@@ -71,7 +78,25 @@ function handleDeepLink(url: string): void {
 
 // --- Window creation -----------------------------------------------------
 
+// Tracks the OS-preferred language as last seen by the running process.
+// Updated on each window-focus check so we can emit a `locale:system-changed`
+// event to the renderer when the user changes their OS language without
+// quitting the app — without restart, app.getPreferredSystemLanguages()
+// would still report the boot value forever.
+let lastKnownSystemLocale = "en";
+
+function getSystemLocale(): string {
+  return app.getPreferredSystemLanguages()[0] ?? "en";
+}
+
 function createWindow(): void {
+  // Pass the OS-preferred language to the renderer via additionalArguments
+  // instead of a sync IPC call. process.argv is available to the preload
+  // script before the first network request, so the renderer's i18next
+  // instance can initialize with the right locale on the very first paint.
+  const systemLocale = getSystemLocale();
+  lastKnownSystemLocale = systemLocale;
+
   mainWindow = new BrowserWindow({
     width: 1280,
     height: 800,
@@ -88,6 +113,7 @@ function createWindow(): void {
       preload: join(__dirname, "../preload/index.js"),
       sandbox: false,
       webSecurity: false,
+      additionalArguments: [`--multica-locale=${systemLocale}`],
     },
   });
 
@@ -105,11 +131,39 @@ function createWindow(): void {
     mainWindow?.show();
   });
 
+  // Detect OS language changes while the app is running. Electron has no
+  // dedicated event for this on any platform, so we poll on focus regain —
+  // catches the common case where users switch System Settings → Language
+  // and bring the app back. The renderer decides whether to act (it ignores
+  // the signal when the user has an explicit Settings choice).
+  mainWindow.on("focus", () => {
+    const current = getSystemLocale();
+    if (current === lastKnownSystemLocale) return;
+    lastKnownSystemLocale = current;
+    mainWindow?.webContents.send("locale:system-changed", current);
+  });
+
   mainWindow.webContents.setWindowOpenHandler((details) => {
     openExternalSafely(details.url);
     return { action: "deny" };
   });
 
+  // Prevent Cmd+R / Ctrl+R / Shift+Cmd+R / Shift+Ctrl+R / F5 from
+  // reloading the page. In a desktop app an accidental reload destroys
+  // in-memory state (tabs, drafts, WS connections) with no URL bar to
+  // navigate back. DevTools refresh (via the DevTools UI) still works.
+  mainWindow.webContents.on("before-input-event", (_event, input) => {
+    if (input.type !== "keyDown") return;
+    const cmdOrCtrl =
+      process.platform === "darwin" ? input.meta : input.control;
+    if (
+      (cmdOrCtrl && input.key.toLowerCase() === "r") ||
+      input.key === "F5"
+    ) {
+      _event.preventDefault();
+    }
+  });
+
   installContextMenu(mainWindow.webContents);
 
   if (is.dev && process.env["ELECTRON_RENDERER_URL"]) {
@@ -170,7 +224,25 @@ if (!gotTheLock) {
     if (deepLinkUrl) handleDeepLink(deepLinkUrl);
   });
 
-  app.whenReady().then(() => {
+  app.whenReady().then(async () => {
+    const viteEnv = import.meta.env as ImportMetaEnv & {
+      readonly VITE_API_URL?: string;
+      readonly VITE_WS_URL?: string;
+      readonly VITE_APP_URL?: string;
+    };
+
+    runtimeConfigResult = await loadRuntimeConfig({
+      isDev: is.dev,
+      // electron-vite exposes VITE_* on import.meta.env for the main process;
+      // keep dev URL overrides on the same source the renderer used before
+      // runtime config moved endpoint resolution into main/preload.
+      env: {
+        apiUrl: viteEnv.VITE_API_URL,
+        wsUrl: viteEnv.VITE_WS_URL,
+        appUrl: viteEnv.VITE_APP_URL,
+      },
+    });
+
     electronApp.setAppUserModelId(
       is.dev ? "ai.multica.desktop.dev" : "ai.multica.desktop",
     );
@@ -203,7 +275,14 @@ if (!gotTheLock) {
     ipcMain.on("app:get-info", (event) => {
       const p = process.platform;
       const os = p === "darwin" ? "macos" : p === "win32" ? "windows" : p === "linux" ? "linux" : "unknown";
-      event.returnValue = { version: app.getVersion(), os };
+      event.returnValue = { version: getAppVersion(), os };
+    });
+
+    // Sync IPC: preload exposes the validated runtime config before renderer
+    // boot. If desktop.json exists but is invalid, renderer receives the
+    // blocking error and must not silently fall back to the cloud defaults.
+    ipcMain.on("runtime-config:get", (event) => {
+      event.returnValue = runtimeConfigResult;
     });
 
     // IPC: toggle immersive mode — hides the macOS traffic lights so full-screen

apps/desktop/src/main/runtime-config-loader.test.ts

@@ -0,0 +1,90 @@
+import { mkdtemp, writeFile } from "fs/promises";
+import { join } from "path";
+import { tmpdir } from "os";
+import { describe, expect, it } from "vitest";
+import { loadRuntimeConfig } from "./runtime-config-loader";
+
+describe("loadRuntimeConfig", () => {
+  it("uses dev env and ignores desktop.json during electron-vite dev", async () => {
+    const dir = await mkdtemp(join(tmpdir(), "multica-desktop-config-"));
+    const configPath = join(dir, "desktop.json");
+    await writeFile(
+      configPath,
+      JSON.stringify({ schemaVersion: 1, apiUrl: "https://prod.example.com" }),
+    );
+
+    await expect(
+      loadRuntimeConfig({
+        isDev: true,
+        configPath,
+        env: {
+          apiUrl: "http://localhost:8080",
+          wsUrl: "ws://localhost:8080/ws",
+          appUrl: "http://localhost:3000",
+        },
+      }),
+    ).resolves.toEqual({
+      ok: true,
+      config: {
+        schemaVersion: 1,
+        apiUrl: "http://localhost:8080",
+        wsUrl: "ws://localhost:8080/ws",
+        appUrl: "http://localhost:3000",
+      },
+    });
+  });
+
+  it("uses cloud defaults when packaged config is absent", async () => {
+    const dir = await mkdtemp(join(tmpdir(), "multica-desktop-config-"));
+    await expect(
+      loadRuntimeConfig({
+        isDev: false,
+        configPath: join(dir, "missing.json"),
+        env: {},
+      }),
+    ).resolves.toEqual({
+      ok: true,
+      config: {
+        schemaVersion: 1,
+        apiUrl: "https://api.multica.ai",
+        wsUrl: "wss://api.multica.ai/ws",
+        appUrl: "https://multica.ai",
+      },
+    });
+  });
+
+  it("parses a valid packaged desktop.json", async () => {
+    const dir = await mkdtemp(join(tmpdir(), "multica-desktop-config-"));
+    const configPath = join(dir, "desktop.json");
+    await writeFile(
+      configPath,
+      JSON.stringify({ schemaVersion: 1, apiUrl: "https://api.example.com" }),
+    );
+
+    await expect(
+      loadRuntimeConfig({ isDev: false, configPath, env: {} }),
+    ).resolves.toEqual({
+      ok: true,
+      config: {
+        schemaVersion: 1,
+        apiUrl: "https://api.example.com",
+        wsUrl: "wss://api.example.com/ws",
+        appUrl: "https://example.com",
+      },
+    });
+  });
+
+  it("fails closed when packaged desktop.json is invalid", async () => {
+    const dir = await mkdtemp(join(tmpdir(), "multica-desktop-config-"));
+    const configPath = join(dir, "desktop.json");
+    await writeFile(configPath, "{");
+
+    const result = await loadRuntimeConfig({ isDev: false, configPath, env: {} });
+
+    expect(result.ok).toBe(false);
+    if (!result.ok) {
+      expect(result.error.message).toContain(configPath);
+      expect(result.error.message).toContain("Invalid desktop runtime config JSON");
+    }
+  });
+});

apps/desktop/src/main/runtime-config-loader.ts

@@ -0,0 +1,60 @@
+import { app } from "electron";
+import { readFile } from "fs/promises";
+import { join } from "path";
+import {
+  DEFAULT_RUNTIME_CONFIG,
+  parseRuntimeConfig,
+  runtimeConfigFromDevEnv,
+  type RuntimeConfig,
+  type RuntimeConfigEnv,
+  type RuntimeConfigResult,
+} from "../shared/runtime-config";
+
+export async function loadRuntimeConfig(options: {
+  isDev: boolean;
+  env: RuntimeConfigEnv;
+  configPath?: string;
+}): Promise<RuntimeConfigResult> {
+  if (options.isDev) {
+    try {
+      return { ok: true, config: runtimeConfigFromDevEnv(options.env) };
+    } catch (err) {
+      return { ok: false, error: { message: errorMessage(err) } };
+    }
+  }
+
+  const configPath = options.configPath ?? desktopConfigPath();
+  try {
+    const raw = await readFile(configPath, "utf-8");
+    return { ok: true, config: parseRuntimeConfig(raw) };
+  } catch (err) {
+    if (isMissingFileError(err)) {
+      return { ok: true, config: { ...DEFAULT_RUNTIME_CONFIG } };
+    }
+    return {
+      ok: false,
+      error: {
+        message: `Invalid ${configPath}: ${errorMessage(err)}`,
+      },
+    };
+  }
+}
+
+export function desktopConfigPath(): string {
+  return join(app.getPath("home"), ".multica", "desktop.json");
+}
+
+function isMissingFileError(err: unknown): boolean {
+  return Boolean(
+    err &&
+      typeof err === "object" &&
+      "code" in err &&
+      (err as NodeJS.ErrnoException).code === "ENOENT",
+  );
+}
+
+function errorMessage(err: unknown): string {
+  return err instanceof Error ? err.message : String(err);
+}
+
+export type { RuntimeConfig, RuntimeConfigResult };

apps/desktop/src/preload/index.d.ts

@@ -1,4 +1,5 @@
 import { ElectronAPI } from "@electron-toolkit/preload";
+import type { RuntimeConfigResult } from "../shared/runtime-config";
 
 interface DesktopAPI {
   /** App version + normalized OS, captured synchronously at preload time. */
@@ -6,6 +7,12 @@ interface DesktopAPI {
     version: string;
     os: "macos" | "windows" | "linux" | "unknown";
   };
+  /** OS-preferred locale (BCP 47) injected by main via additionalArguments. */
+  systemLocale: string;
+  /** Subscribe to OS language changes detected after boot. Returns an unsubscribe function. */
+  onSystemLocaleChanged: (callback: (locale: string) => void) => () => void;
+  /** Validated runtime endpoint config, or a blocking config error. */
+  runtimeConfig: RuntimeConfigResult;
   /** Listen for auth token delivered via deep link. Returns an unsubscribe function. */
   onAuthToken: (callback: (token: string) => void) => () => void;
   /** Listen for invitation IDs delivered via deep link. Returns an unsubscribe function. */

apps/desktop/src/preload/index.ts

@@ -1,5 +1,6 @@
 import { contextBridge, ipcRenderer } from "electron";
 import { electronAPI } from "@electron-toolkit/preload";
+import type { RuntimeConfigResult } from "../shared/runtime-config";
 
 // Synchronously fetch app metadata from main at preload time so the renderer
 // can pass it into CoreProvider during the initial render — the alternative
@@ -21,12 +22,53 @@ function fetchAppInfo(): { version: string; os: "macos" | "windows" | "linux" |
   return { version: "unknown", os };
 }
 
+function fetchRuntimeConfig(): RuntimeConfigResult {
+  try {
+    const result = ipcRenderer.sendSync("runtime-config:get") as RuntimeConfigResult | undefined;
+    if (result && typeof result === "object" && "ok" in result) return result;
+  } catch (err) {
+    return {
+      ok: false,
+      error: {
+        message: err instanceof Error ? err.message : String(err),
+      },
+    };
+  }
+  return { ok: false, error: { message: "Runtime config unavailable" } };
+}
+
 const appInfo = fetchAppInfo();
+const runtimeConfig = fetchRuntimeConfig();
+
+// Read the OS-preferred locale that main injected via additionalArguments.
+// Zero IPC, zero blocking — process.argv is populated before preload runs.
+function fetchSystemLocale(): string {
+  const arg = process.argv.find((a) => a.startsWith("--multica-locale="));
+  return arg?.split("=")[1] ?? "en";
+}
+
+const systemLocale = fetchSystemLocale();
 
 const desktopAPI = {
   /** App version + normalized OS. Read once at preload time so the renderer
    *  can use it synchronously when initializing the API client. */
   appInfo,
+  /** OS-preferred locale (BCP 47), passed from main via additionalArguments.
+   *  Used by the renderer's LocaleAdapter as the system-preference signal. */
+  systemLocale,
+  /** Subscribe to OS language changes detected after boot. The renderer
+   *  decides whether to act (no-op when the user has an explicit Settings
+   *  choice). Returns an unsubscribe function. */
+  onSystemLocaleChanged: (callback: (locale: string) => void) => {
+    const handler = (_event: Electron.IpcRendererEvent, locale: string) =>
+      callback(locale);
+    ipcRenderer.on("locale:system-changed", handler);
+    return () => {
+      ipcRenderer.removeListener("locale:system-changed", handler);
+    };
+  },
+  /** Validated runtime endpoint config, or a blocking config error. */
+  runtimeConfig,
   /** Listen for auth token delivered via deep link */
   onAuthToken: (callback: (token: string) => void) => {
     const handler = (_event: Electron.IpcRendererEvent, token: string) =>

apps/desktop/src/renderer/src/App.tsx

@@ -1,6 +1,7 @@
 import { useEffect, useLayoutEffect, useMemo, useRef, useState } from "react";
 import { useQuery, useQueryClient } from "@tanstack/react-query";
 import { CoreProvider } from "@multica/core/platform";
+import { pickLocale } from "@multica/core/i18n";
 import { useAuthStore } from "@multica/core/auth";
 import { workspaceKeys, workspaceListOptions } from "@multica/core/workspace/queries";
 import { api } from "@multica/core/api";
@@ -15,6 +16,8 @@ import { UpdateNotification } from "./components/update-notification";
 import { useTabStore } from "./stores/tab-store";
 import { useWindowOverlayStore } from "./stores/window-overlay-store";
 import { useDaemonIPCBridge } from "./platform/daemon-ipc-bridge";
+import { createDesktopLocaleAdapter } from "./platform/i18n-adapter";
+import { RESOURCES } from "@multica/views/locales";
 
 
 function AppContent() {
@@ -30,11 +33,16 @@ function AppContent() {
   // first render.
   const [bootstrapping, setBootstrapping] = useState(false);
 
+  const runtimeConfig = window.desktopAPI.runtimeConfig.ok
+    ? window.desktopAPI.runtimeConfig.config
+    : null;
+
   // Tell the main process which backend URL we talk to, so daemon-manager
   // can pick the matching CLI profile (server_url from ~/.multica config).
   useEffect(() => {
-    window.daemonAPI.setTargetApiUrl(DAEMON_TARGET_API_URL);
-  }, []);
+    if (!runtimeConfig) return;
+    window.daemonAPI.setTargetApiUrl(runtimeConfig.apiUrl);
+  }, [runtimeConfig]);
 
   // Listen for invite IDs delivered via deep link (multica://invite/<id>).
   // We open the overlay regardless of login state — if the user isn't logged
@@ -110,21 +118,58 @@ function AppContent() {
     : undefined;
   useDaemonIPCBridge(activeWsId);
 
-  // Onboarding and zero-workspace both resolve to an overlay, but
-  // onboarding wins: a user who hasn't completed it gets the onboarding
-  // overlay regardless of how many workspaces already exist.
+  // Pre-workspace overlay routing for desktop. Mirrors the web entry-point
+  // judgment in callback / login:
+  //   un-onboarded:
+  //     pending invites on email → /invitations overlay
+  //     no invites               → /onboarding overlay
+  //   already onboarded:
+  //     zero workspaces          → /workspaces/new overlay
+  //     ≥1 workspaces            → no overlay, fall through to dashboard
+  //
+  // The "un-onboarded but in workspace" state is now physically impossible
+  // because backend transactions atomically set onboarded_at when a user
+  // joins the `member` table. Anyone with workspaces is by definition
+  // onboarded.
   useEffect(() => {
-    if (!user || !workspaceListFetched) return;
+    if (!user || !workspaceListFetched) return undefined;
     const { overlay, open } = useWindowOverlayStore.getState();
-    if (overlay) return;
+    if (overlay) return undefined;
+    if (wsCount > 0) return undefined;
     if (!hasOnboarded) {
-      open({ type: "onboarding" });
-      return;
+      // Look up pending invitations by email. Network blip is non-fatal —
+      // fall through to onboarding so the user isn't stuck on a blank
+      // window. The sidebar's pending-invitations dropdown will surface
+      // missed invites later once they're onboarded.
+      let cancelled = false;
+      void api
+        .listMyInvitations()
+        .then((invites) => {
+          if (cancelled) return;
+          const { overlay: latestOverlay, open: latestOpen } =
+            useWindowOverlayStore.getState();
+          if (latestOverlay) return;
+          if (invites.length > 0) {
+            qc.setQueryData(workspaceKeys.myInvitations(), invites);
+            latestOpen({ type: "invitations" });
+          } else {
+            latestOpen({ type: "onboarding" });
+          }
+        })
+        .catch(() => {
+          if (cancelled) return;
+          const { overlay: latestOverlay, open: latestOpen } =
+            useWindowOverlayStore.getState();
+          if (latestOverlay) return;
+          latestOpen({ type: "onboarding" });
+        });
+      return () => {
+        cancelled = true;
+      };
     }
-    if (wsCount === 0) {
-      open({ type: "new-workspace" });
-    }
-  }, [user, workspaceListFetched, wsCount, workspaces, hasOnboarded]);
+    open({ type: "new-workspace" });
+    return undefined;
+  }, [user, workspaceListFetched, wsCount, workspaces, hasOnboarded, qc]);
 
   // Validate persisted tab state against the current user's workspace list,
   // and pick an active workspace if none is set. Runs in useLayoutEffect
@@ -189,9 +234,21 @@ function AppContent() {
   );
 }
 
-// Backend the daemon should connect to — same URL the renderer talks to.
-const DAEMON_TARGET_API_URL =
-  import.meta.env.VITE_API_URL || "http://localhost:8080";
+function BlockingRuntimeConfigError({ message }: { message: string }) {
+  return (
+    <div className="flex h-screen items-center justify-center bg-background p-8 text-foreground">
+      <div className="max-w-xl rounded-lg border bg-card p-6 shadow-sm">
+        <h1 className="text-lg font-semibold">Desktop configuration error</h1>
+        <p className="mt-3 text-sm text-muted-foreground">
+          Multica Desktop could not load <code>~/.multica/desktop.json</code>. Fix or remove the file and restart the app.
+        </p>
+        <pre className="mt-4 whitespace-pre-wrap rounded-md bg-muted p-3 text-xs text-muted-foreground">
+          {message}
+        </pre>
+      </div>
+    </div>
+  );
+}
 
 // On logout, wipe desktop-only in-memory state and stop the daemon so that
 // a subsequent login as a different user never inherits the previous user's
@@ -215,22 +272,61 @@ async function handleDaemonLogout() {
 
 export default function App() {
   const { version, os } = window.desktopAPI.appInfo;
+  const systemLocale = window.desktopAPI.systemLocale;
+  const runtimeConfigResult = window.desktopAPI.runtimeConfig;
   // Stable identity reference so downstream effects (WS reconnect) don't
   // tear down on every parent render.
   const identity = useMemo(
     () => ({ platform: "desktop", version, os }),
     [version, os],
   );
+  // Locale resolution happens once at app boot. Switching language goes
+  // through window.location.reload() to avoid hydration mismatch.
+  const localeAdapter = useMemo(
+    () => createDesktopLocaleAdapter(systemLocale),
+    [systemLocale],
+  );
+  const locale = useMemo(() => pickLocale(localeAdapter), [localeAdapter]);
+  const resources = useMemo(
+    () => ({ [locale]: RESOURCES[locale] }),
+    [locale],
+  );
+
+  // React to OS-level language changes detected by main on focus regain.
+  // Only act when the user is following the system signal (no explicit
+  // Settings choice) — otherwise their preference wins. Cross-device sync
+  // for the explicit-choice case is handled inside CoreProvider.
+  useEffect(() => {
+    return window.desktopAPI.onSystemLocaleChanged((nextSystemLocale) => {
+      if (localeAdapter.getUserChoice()) return;
+      const next = pickLocale({
+        ...localeAdapter,
+        getSystemPreferences: () =>
+          nextSystemLocale ? [nextSystemLocale] : [],
+      });
+      if (next === locale) return;
+      localeAdapter.persist(next);
+      window.location.reload();
+    });
+  }, [localeAdapter, locale]);
+
   return (
     <ThemeProvider>
-      <CoreProvider
-        apiBaseUrl={import.meta.env.VITE_API_URL || "http://localhost:8080"}
-        wsUrl={import.meta.env.VITE_WS_URL || "ws://localhost:8080/ws"}
-        onLogout={handleDaemonLogout}
-        identity={identity}
-      >
-        <AppContent />
-      </CoreProvider>
+      {runtimeConfigResult.ok ? (
+        <CoreProvider
+          apiBaseUrl={runtimeConfigResult.config.apiUrl}
+          wsUrl={runtimeConfigResult.config.wsUrl}
+          onLogout={handleDaemonLogout}
+          identity={identity}
+          locale={locale}
+          resources={resources}
+          localeAdapter={localeAdapter}
+        >
+          <AppContent />
+        </CoreProvider>
+      ) : (
+        <BlockingRuntimeConfigError message={runtimeConfigResult.error.message} />
+      )}
       <Toaster />
       <UpdateNotification />
     </ThemeProvider>

apps/desktop/src/renderer/src/components/pageview-tracker.test.tsx

@@ -0,0 +1,243 @@
+import { describe, expect, it, vi, beforeEach } from "vitest";
+import { render } from "@testing-library/react";
+
+// vi.hoisted shared state — every store mock reads the same object so each
+// test can mutate it then re-render to drive the tracker.
+const state = vi.hoisted(() => ({
+  user: null as { id: string } | null,
+  overlay: null as { type: string; invitationId?: string } | null,
+  activeWorkspaceSlug: null as string | null,
+  byWorkspace: {} as Record<
+    string,
+    { activeTabId: string; tabs: { id: string; path: string }[] }
+  >,
+  capturePageview: vi.fn<(path?: string) => void>(),
+}));
+
+vi.mock("@multica/core/analytics", () => ({
+  capturePageview: state.capturePageview,
+}));
+
+// Auth store — single selector pattern (`s => s.user`).
+vi.mock("@multica/core/auth", () => {
+  const useAuthStore = (selector: (s: typeof state) => unknown) =>
+    selector(state);
+  return { useAuthStore };
+});
+
+// Window overlay store — same shape.
+vi.mock("@/stores/window-overlay-store", () => {
+  const useWindowOverlayStore = (selector: (s: typeof state) => unknown) =>
+    selector(state);
+  return { useWindowOverlayStore };
+});
+
+// Tab store — selectors read activeWorkspaceSlug + byWorkspace. Also expose
+// getState() for the seed pass and the helpers the tracker imports
+// (useActiveTabIdentity, getActiveTab) so we don't have to re-import them
+// from the real store inside a mocked module.
+vi.mock("@/stores/tab-store", () => {
+  const useTabStore = Object.assign(
+    (selector: (s: typeof state) => unknown) => selector(state),
+    { getState: () => state },
+  );
+  const getActiveTab = (s: typeof state) => {
+    const slug = s.activeWorkspaceSlug;
+    if (!slug) return null;
+    const group = s.byWorkspace[slug];
+    if (!group) return null;
+    return group.tabs.find((t) => t.id === group.activeTabId) ?? null;
+  };
+  const useActiveTabIdentity = () => ({
+    slug: state.activeWorkspaceSlug,
+    tabId: state.activeWorkspaceSlug
+      ? (state.byWorkspace[state.activeWorkspaceSlug]?.activeTabId ?? null)
+      : null,
+  });
+  return { useTabStore, getActiveTab, useActiveTabIdentity };
+});
+
+import { PageviewTracker } from "./pageview-tracker";
+
+function reset() {
+  state.user = { id: "u1" };
+  state.overlay = null;
+  state.activeWorkspaceSlug = null;
+  state.byWorkspace = {};
+  state.capturePageview.mockClear();
+}
+
+beforeEach(() => {
+  reset();
+});
+
+describe("PageviewTracker", () => {
+  it("suppresses pageview when switching to a previously-visible tab on its existing path", () => {
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tA",
+        tabs: [
+          { id: "tA", path: "/acme/issues" },
+          { id: "tB", path: "/acme/inbox" },
+        ],
+      },
+    };
+    state.activeWorkspaceSlug = "acme";
+
+    const { rerender } = render(<PageviewTracker />);
+    // Initial mount on tA — seeded as observed, no pageview because both
+    // tabs were already in the persisted store before the tracker mounted.
+    expect(state.capturePageview).not.toHaveBeenCalled();
+
+    // Switch to tB (already-known tab on its already-known path).
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tB",
+        tabs: [
+          { id: "tA", path: "/acme/issues" },
+          { id: "tB", path: "/acme/inbox" },
+        ],
+      },
+    };
+    rerender(<PageviewTracker />);
+    expect(state.capturePageview).not.toHaveBeenCalled();
+
+    // Switch back to tA — still no pageview.
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tA",
+        tabs: [
+          { id: "tA", path: "/acme/issues" },
+          { id: "tB", path: "/acme/inbox" },
+        ],
+      },
+    };
+    rerender(<PageviewTracker />);
+    expect(state.capturePageview).not.toHaveBeenCalled();
+  });
+
+  it("fires pageview when a new tab is opened (openInNewTab / addTab)", () => {
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tA",
+        tabs: [{ id: "tA", path: "/acme/issues" }],
+      },
+    };
+    state.activeWorkspaceSlug = "acme";
+
+    const { rerender } = render(<PageviewTracker />);
+    state.capturePageview.mockClear();
+
+    // Simulate openInNewTab("/acme/agents") → new tab tC added and activated.
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tC",
+        tabs: [
+          { id: "tA", path: "/acme/issues" },
+          { id: "tC", path: "/acme/agents" },
+        ],
+      },
+    };
+    rerender(<PageviewTracker />);
+
+    expect(state.capturePageview).toHaveBeenCalledTimes(1);
+    expect(state.capturePageview).toHaveBeenCalledWith("/acme/agents");
+  });
+
+  it("fires pageview when switchWorkspace opens a new path in another workspace", () => {
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tA",
+        tabs: [{ id: "tA", path: "/acme/issues" }],
+      },
+    };
+    state.activeWorkspaceSlug = "acme";
+
+    const { rerender } = render(<PageviewTracker />);
+    state.capturePageview.mockClear();
+
+    // Cross-workspace navigation: switchWorkspace("butter", "/butter/inbox")
+    // creates a fresh tab in the destination workspace and makes it active.
+    state.byWorkspace = {
+      acme: { activeTabId: "tA", tabs: [{ id: "tA", path: "/acme/issues" }] },
+      butter: {
+        activeTabId: "tD",
+        tabs: [{ id: "tD", path: "/butter/inbox" }],
+      },
+    };
+    state.activeWorkspaceSlug = "butter";
+    rerender(<PageviewTracker />);
+
+    expect(state.capturePageview).toHaveBeenCalledTimes(1);
+    expect(state.capturePageview).toHaveBeenCalledWith("/butter/inbox");
+  });
+
+  it("fires pageview on intra-tab navigation (path changes for the same tabId)", () => {
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tA",
+        tabs: [{ id: "tA", path: "/acme/issues" }],
+      },
+    };
+    state.activeWorkspaceSlug = "acme";
+
+    const { rerender } = render(<PageviewTracker />);
+    state.capturePageview.mockClear();
+
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tA",
+        tabs: [{ id: "tA", path: "/acme/issues/123" }],
+      },
+    };
+    rerender(<PageviewTracker />);
+
+    expect(state.capturePageview).toHaveBeenCalledTimes(1);
+    expect(state.capturePageview).toHaveBeenCalledWith("/acme/issues/123");
+  });
+
+  it("fires overlay and login pageviews and suppresses re-entry into the same tab afterward", () => {
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tA",
+        tabs: [{ id: "tA", path: "/acme/issues" }],
+      },
+    };
+    state.activeWorkspaceSlug = "acme";
+
+    const { rerender } = render(<PageviewTracker />);
+    state.capturePageview.mockClear();
+
+    // Open onboarding overlay.
+    state.overlay = { type: "onboarding" };
+    rerender(<PageviewTracker />);
+    expect(state.capturePageview).toHaveBeenLastCalledWith("/onboarding");
+
+    // Close overlay back to the tab — the tab is already observed on
+    // /acme/issues so this is a re-activation, no pageview.
+    state.capturePageview.mockClear();
+    state.overlay = null;
+    rerender(<PageviewTracker />);
+    expect(state.capturePageview).not.toHaveBeenCalled();
+
+    // Logout fires /login.
+    state.user = null;
+    rerender(<PageviewTracker />);
+    expect(state.capturePageview).toHaveBeenLastCalledWith("/login");
+  });
+
+  it("suppresses on initial mount when the active tab was restored from persistence", () => {
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tA",
+        tabs: [{ id: "tA", path: "/acme/issues" }],
+      },
+    };
+    state.activeWorkspaceSlug = "acme";
+
+    render(<PageviewTracker />);
+    // Restored tab — seeded, treated as a re-activation.
+    expect(state.capturePageview).not.toHaveBeenCalled();
+  });
+});
+

apps/desktop/src/renderer/src/components/pageview-tracker.tsx

@@ -1,11 +1,17 @@
-import { useEffect } from "react";
+import { useEffect, useRef } from "react";
 import { capturePageview } from "@multica/core/analytics";
 import { useAuthStore } from "@multica/core/auth";
-import { useTabStore } from "@/stores/tab-store";
+import {
+  getActiveTab,
+  useActiveTabIdentity,
+  useTabStore,
+} from "@/stores/tab-store";
 import { useWindowOverlayStore, type WindowOverlay } from "@/stores/window-overlay-store";
 
 /**
- * Fires a PostHog $pageview whenever the user's visible surface changes.
+ * Fires a PostHog $pageview whenever the user's visible surface changes,
+ * EXCEPT for re-activations of an already-known tab on its already-known
+ * path.
  *
  * Desktop has three layers that can own the visible page:
  *
@@ -17,10 +23,18 @@ import { useWindowOverlayStore, type WindowOverlay } from "@/stores/window-overl
  *   3. Otherwise → the active tab's path (workspace-scoped, e.g.
  *      `/acme/issues/123`). Kept in sync by `useTabRouterSync`.
  *
- * The overlay takes precedence over the tab path because it is visually in
- * front of the tab system; the logged-out state shadows both because the
- * shell doesn't render at all yet. This keeps the `$pageview` stream aligned
- * with what the user actually sees.
+ * Tab-switch suppression: re-activating an already-open tab surfaces a
+ * previously-visited path under a `(workspace, tabId)` we have already
+ * seen — the pageview was emitted when the user originally navigated
+ * there, so re-emitting on every switch just inflates PostHog billing
+ * without adding signal (real-data audit: desktop tab switches were
+ * ~50% of all `$pageview` events).
+ *
+ * Newly opened tabs (`openInNewTab`, `addTab`) and cross-workspace
+ * `switchWorkspace(slug, path)` to a previously-unseen tab still fire,
+ * because their key is not in the observed map yet. The map is seeded
+ * from the persisted tab store on first render so tabs restored from a
+ * previous session don't all re-emit on first activation.
  *
  * PostHog's `capture_pageview: true` auto-capture is intentionally off (see
  * `initAnalytics`) so this component owns the event shape, matching the web
@@ -29,34 +43,75 @@ import { useWindowOverlayStore, type WindowOverlay } from "@/stores/window-overl
 export function PageviewTracker() {
   const user = useAuthStore((s) => s.user);
   const overlay = useWindowOverlayStore((s) => s.overlay);
-  const activeTabPath = useTabStore((s) => {
-    const slug = s.activeWorkspaceSlug;
-    if (!slug) return null;
-    const group = s.byWorkspace[slug];
-    if (!group) return null;
-    return group.tabs.find((t) => t.id === group.activeTabId)?.path ?? null;
-  });
+  const { slug: activeWorkspaceSlug, tabId: activeTabId } = useActiveTabIdentity();
+  const activeTabPath = useTabStore((s) => getActiveTab(s)?.path ?? null);
 
-  const path = resolvePath(user, overlay, activeTabPath);
+  // (slug:tabId) → last path observed while that tab was visible. Lets us
+  // tell "re-activating a tab on a path we already saw" (suppress) apart
+  // from "newly opened tab" or "intra-tab navigation" (fire). Seeded
+  // synchronously on first render from the persisted tab store so
+  // session-restored tabs don't re-emit on first click.
+  const observedTabsRef = useRef<Map<string, string> | null>(null);
+  if (observedTabsRef.current === null) {
+    const seed = new Map<string, string>();
+    for (const [slug, group] of Object.entries(useTabStore.getState().byWorkspace)) {
+      for (const tab of group.tabs) {
+        seed.set(`${slug}:${tab.id}`, tab.path);
+      }
+    }
+    observedTabsRef.current = seed;
+  }
+  const lastSurfaceRef = useRef<{
+    kind: "login" | "overlay" | "tab" | null;
+    key: string | null;
+    path: string | null;
+  }>({ kind: null, key: null, path: null });
 
   useEffect(() => {
-    if (!path) return;
+    let kind: "login" | "overlay" | "tab";
+    let path: string;
+    let key: string | null = null;
+
+    if (!user) {
+      kind = "login";
+      path = "/login";
+    } else if (overlay) {
+      kind = "overlay";
+      path = overlayPath(overlay);
+    } else if (activeTabPath && activeTabId && activeWorkspaceSlug) {
+      kind = "tab";
+      key = `${activeWorkspaceSlug}:${activeTabId}`;
+      path = activeTabPath;
+    } else {
+      return;
+    }
+
+    const observed = observedTabsRef.current!;
+    const last = lastSurfaceRef.current;
+    const next = { kind, key, path };
+
+    if (kind === "tab" && key !== null) {
+      const knownPath = observed.get(key);
+      const isReactivation =
+        last.key !== key && knownPath !== undefined && knownPath === path;
+      observed.set(key, path);
+      if (isReactivation) {
+        lastSurfaceRef.current = next;
+        return;
+      }
+    }
+
+    const unchanged =
+      last.kind === kind && last.key === key && last.path === path;
+    if (unchanged) return;
+
     capturePageview(path);
-  }, [path]);
+    lastSurfaceRef.current = next;
+  }, [user, overlay, activeWorkspaceSlug, activeTabId, activeTabPath]);
 
   return null;
 }
 
-function resolvePath(
-  user: unknown,
-  overlay: WindowOverlay | null,
-  activeTabPath: string | null,
-): string | null {
-  if (!user) return "/login";
-  if (overlay) return overlayPath(overlay);
-  return activeTabPath;
-}
-
 function overlayPath(overlay: WindowOverlay): string {
   switch (overlay.type) {
     case "new-workspace":
@@ -65,5 +120,7 @@ function overlayPath(overlay: WindowOverlay): string {
       return "/onboarding";
     case "invite":
       return `/invite/${overlay.invitationId}`;
+    case "invitations":
+      return "/invitations";
   }
 }

apps/desktop/src/renderer/src/components/updates-settings-tab.tsx

@@ -5,12 +5,13 @@ import { Button } from "@multica/ui/components/ui/button";
 type CheckState =
   | { status: "idle" }
   | { status: "checking" }
-  | { status: "up-to-date"; currentVersion: string }
+  | { status: "up-to-date" }
   | { status: "available"; latestVersion: string }
   | { status: "error"; message: string };
 
 export function UpdatesSettingsTab() {
   const [state, setState] = useState<CheckState>({ status: "idle" });
+  const currentVersion = window.desktopAPI.appInfo.version;
 
   const handleCheck = useCallback(async () => {
     setState({ status: "checking" });
@@ -22,7 +23,7 @@ export function UpdatesSettingsTab() {
     setState(
       result.available
         ? { status: "available", latestVersion: result.latestVersion }
-        : { status: "up-to-date", currentVersion: result.currentVersion },
+        : { status: "up-to-date" },
     );
   }, []);
 
@@ -35,6 +36,15 @@ export function UpdatesSettingsTab() {
       </p>
 
       <div className="mt-6 divide-y">
+        <div className="flex items-center justify-between gap-6 py-4">
+          <div className="min-w-0">
+            <p className="text-sm font-medium">Current version</p>
+            <p className="text-sm text-muted-foreground mt-0.5 font-mono">
+              v{currentVersion}
+            </p>
+          </div>
+        </div>
+
         <div className="flex items-start justify-between gap-6 py-4">
           <div className="min-w-0">
             <p className="text-sm font-medium">Check for updates</p>
@@ -45,7 +55,7 @@ export function UpdatesSettingsTab() {
             {state.status === "up-to-date" && (
               <p className="text-sm text-muted-foreground mt-2 inline-flex items-center gap-1.5">
                 <Check className="size-3.5 text-success" />
-                You&apos;re on the latest version (v{state.currentVersion}).
+                You&apos;re on the latest version.
               </p>
             )}
             {state.status === "available" && (

apps/desktop/src/renderer/src/components/window-overlay.tsx

@@ -1,6 +1,7 @@
 import { useQuery } from "@tanstack/react-query";
 import { NewWorkspacePage } from "@multica/views/workspace/new-workspace-page";
 import { InvitePage } from "@multica/views/invite";
+import { InvitationsPage } from "@multica/views/invitations";
 import { OnboardingFlow } from "@multica/views/onboarding";
 import { useNavigation } from "@multica/views/navigation";
 import { paths } from "@multica/core/paths";
@@ -58,6 +59,7 @@ function WindowOverlayInner() {
           onBack={onBack}
         />
       )}
+      {overlay.type === "invitations" && <InvitationsPage />}
       {overlay.type === "onboarding" && (
         <OnboardingFlow
           onComplete={(ws) => {

apps/desktop/src/renderer/src/pages/issue-detail-page.tsx

@@ -1,6 +1,7 @@
 import { useParams } from "react-router-dom";
 import { useQuery } from "@tanstack/react-query";
 import { IssueDetail } from "@multica/views/issues/components";
+import { ErrorBoundary } from "@multica/ui/components/common/error-boundary";
 import { useWorkspaceId } from "@multica/core/hooks";
 import { issueDetailOptions } from "@multica/core/issues/queries";
 import { useDocumentTitle } from "@/hooks/use-document-title";
@@ -13,5 +14,9 @@ export function IssueDetailPage() {
   useDocumentTitle(issue ? `${issue.identifier}: ${issue.title}` : "Issue");
 
   if (!id) return null;
-  return <IssueDetail issueId={id} />;
+  return (
+    <ErrorBoundary resetKeys={[id]}>
+      <IssueDetail issueId={id} />
+    </ErrorBoundary>
+  );
 }

apps/desktop/src/renderer/src/pages/login.tsx

@@ -2,14 +2,23 @@ import { LoginPage } from "@multica/views/auth";
 import { DragStrip } from "@multica/views/platform";
 import { MulticaIcon } from "@multica/ui/components/common/multica-icon";
 
-const WEB_URL = import.meta.env.VITE_APP_URL || "http://localhost:3000";
+function requireRuntimeAppUrl(): string {
+  const runtimeConfig = window.desktopAPI.runtimeConfig;
+  if (!runtimeConfig.ok) {
+    throw new Error(
+      "Invariant violated: DesktopLoginPage rendered before App accepted runtime config",
+    );
+  }
+  return runtimeConfig.config.appUrl;
+}
 
 export function DesktopLoginPage() {
+  const webUrl = requireRuntimeAppUrl();
   const handleGoogleLogin = () => {
     // Open web login page in the default browser with platform=desktop flag.
     // The web callback will redirect back via multica:// deep link with the token.
     window.desktopAPI.openExternal(
-      `${WEB_URL}/login?platform=desktop`,
+      `${webUrl}/login?platform=desktop`,
     );
   };
 

apps/desktop/src/renderer/src/platform/i18n-adapter.ts

@@ -0,0 +1,31 @@
+import type { LocaleAdapter, SupportedLocale } from "@multica/core/i18n";
+
+const STORAGE_KEY = "multica-locale";
+
+// Desktop adapter:
+//   - User choice: localStorage (set by Settings switcher).
+//   - System preference: locale main injected via additionalArguments
+//     (read from preload, exposed on window.desktopAPI.systemLocale).
+//   - Persist: localStorage. The Settings switcher additionally PATCHes
+//     /api/me when logged in so user.language follows the user across devices.
+export function createDesktopLocaleAdapter(systemLocale: string): LocaleAdapter {
+  return {
+    getUserChoice() {
+      try {
+        return window.localStorage.getItem(STORAGE_KEY);
+      } catch {
+        return null;
+      }
+    },
+    getSystemPreferences() {
+      return systemLocale ? [systemLocale] : [];
+    },
+    persist(locale: SupportedLocale) {
+      try {
+        window.localStorage.setItem(STORAGE_KEY, locale);
+      } catch {
+        // Best-effort
+      }
+    },
+  };
+}

apps/desktop/src/renderer/src/platform/navigation.tsx

@@ -15,11 +15,15 @@ import {
 } from "@/stores/tab-store";
 import { useWindowOverlayStore } from "@/stores/window-overlay-store";
 
-// Public web app URL — injected at build time via .env.production. In dev
-// (no VITE_APP_URL set) falls back to the local web dev server so "Copy
-// link" in a dev build yields a URL that points at the running dev
-// frontend, not the prod host. Matches the fallback used in pages/login.tsx.
-const APP_URL = import.meta.env.VITE_APP_URL || "http://localhost:3000";
+function requireRuntimeAppUrl(scope: string): string {
+  const runtimeConfig = window.desktopAPI.runtimeConfig;
+  if (!runtimeConfig.ok) {
+    throw new Error(
+      `Invariant violated: ${scope} rendered before App accepted runtime config`,
+    );
+  }
+  return runtimeConfig.config.appUrl;
+}
 
 /**
  * Extract the leading workspace slug from a path, or null if the path isn't
@@ -61,6 +65,13 @@ function tryRouteToOverlay(path: string, router?: DataRouter): boolean {
     }
     return true;
   }
+  if (path === "/invitations") {
+    overlay.open({ type: "invitations" });
+    if (router && router.state.location.pathname !== "/") {
+      router.navigate("/", { replace: true });
+    }
+    return true;
+  }
   if (path.startsWith("/invite/")) {
     let id = "";
     try {
@@ -109,6 +120,7 @@ export function DesktopNavigationProvider({
 }: {
   children: React.ReactNode;
 }) {
+  const appUrl = requireRuntimeAppUrl("DesktopNavigationProvider");
   // Primitive-only subscriptions so this component doesn't re-render on
   // unrelated store updates (e.g. an inactive tab's router tick). We
   // resolve the active router here only to subscribe once per tab switch.
@@ -179,9 +191,9 @@ export function DesktopNavigationProvider({
         const tabId = store.openTab(path, title ?? path, icon);
         if (tabId) store.setActiveTab(tabId);
       },
-      getShareableUrl: (path: string) => `${APP_URL}${path}`,
+      getShareableUrl: (path: string) => `${appUrl}${path}`,
     }),
-    [location],
+    [appUrl, location],
   );
 
   return <NavigationProvider value={adapter}>{children}</NavigationProvider>;
@@ -204,6 +216,7 @@ export function TabNavigationProvider({
   router: DataRouter;
   children: React.ReactNode;
 }) {
+  const appUrl = requireRuntimeAppUrl("TabNavigationProvider");
   const [location, setLocation] = useState(router.state.location);
 
   useEffect(() => {
@@ -239,9 +252,9 @@ export function TabNavigationProvider({
         const tabId = store.openTab(path, title ?? path, icon);
         if (tabId) store.setActiveTab(tabId);
       },
-      getShareableUrl: (path: string) => `${APP_URL}${path}`,
+      getShareableUrl: (path: string) => `${appUrl}${path}`,
     }),
-    [router, location],
+    [appUrl, router, location],
   );
 
   return <NavigationProvider value={adapter}>{children}</NavigationProvider>;

apps/desktop/src/renderer/src/routes.tsx

@@ -21,6 +21,7 @@ import { DesktopRuntimesPage } from "./components/desktop-runtimes-page";
 import { AgentsPage } from "@multica/views/agents";
 import { InboxPage } from "@multica/views/inbox";
 import { SettingsPage } from "@multica/views/settings";
+import { ErrorBoundary } from "@multica/ui/components/common/error-boundary";
 import { Download, Server } from "lucide-react";
 import { DaemonSettingsTab } from "./components/daemon-settings-tab";
 import { UpdatesSettingsTab } from "./components/updates-settings-tab";
@@ -83,7 +84,15 @@ export const appRoutes: RouteObject[] = [
         element: <WorkspaceRouteLayout />,
         children: [
           { index: true, element: <Navigate to="issues" replace /> },
-          { path: "issues", element: <IssuesPage />, handle: { title: "Issues" } },
+          {
+            path: "issues",
+            element: (
+              <ErrorBoundary>
+                <IssuesPage />
+              </ErrorBoundary>
+            ),
+            handle: { title: "Issues" },
+          },
           {
             path: "issues/:id",
             element: <IssueDetailPage />,

apps/desktop/src/renderer/src/stores/window-overlay-store.ts

@@ -15,6 +15,7 @@ import { create } from "zustand";
 export type WindowOverlay =
   | { type: "new-workspace" }
   | { type: "invite"; invitationId: string }
+  | { type: "invitations" }
   | { type: "onboarding" };
 
 interface WindowOverlayStore {

apps/desktop/src/shared/runtime-config.test.ts

@@ -0,0 +1,151 @@
+import { describe, expect, it } from "vitest";
+import {
+  DEFAULT_RUNTIME_CONFIG,
+  deriveWsUrl,
+  parseRuntimeConfig,
+  runtimeConfigFromDevEnv,
+} from "./runtime-config";
+
+describe("runtime config", () => {
+  it("uses cloud defaults without a desktop.json file", () => {
+    expect(DEFAULT_RUNTIME_CONFIG).toEqual({
+      schemaVersion: 1,
+      apiUrl: "https://api.multica.ai",
+      wsUrl: "wss://api.multica.ai/ws",
+      appUrl: "https://multica.ai",
+    });
+  });
+
+  it("derives https/wss compatible URLs from apiUrl", () => {
+    expect(
+      parseRuntimeConfig(
+        JSON.stringify({
+          schemaVersion: 1,
+          apiUrl: "https://congvc-x99.taila6fa8a.ts.net:18443",
+        }),
+      ),
+    ).toEqual({
+      schemaVersion: 1,
+      apiUrl: "https://congvc-x99.taila6fa8a.ts.net:18443",
+      wsUrl: "wss://congvc-x99.taila6fa8a.ts.net:18443/ws",
+      appUrl: "https://congvc-x99.taila6fa8a.ts.net:18443",
+    });
+  });
+
+  it("strips the leading api. label when deriving appUrl", () => {
+    expect(
+      parseRuntimeConfig(
+        JSON.stringify({ schemaVersion: 1, apiUrl: "https://api.multica.ai" }),
+      ),
+    ).toEqual({
+      schemaVersion: 1,
+      apiUrl: "https://api.multica.ai",
+      wsUrl: "wss://api.multica.ai/ws",
+      appUrl: "https://multica.ai",
+    });
+  });
+
+  it("derives ws for http api URLs", () => {
+    expect(deriveWsUrl("http://localhost:8080")).toBe("ws://localhost:8080/ws");
+  });
+
+  it("accepts explicit appUrl and wsUrl", () => {
+    expect(
+      parseRuntimeConfig(
+        JSON.stringify({
+          schemaVersion: 1,
+          apiUrl: "https://api.example.com/",
+          wsUrl: "wss://ws.example.com/socket/",
+          appUrl: "https://app.example.com/",
+        }),
+      ),
+    ).toEqual({
+      schemaVersion: 1,
+      apiUrl: "https://api.example.com",
+      wsUrl: "wss://ws.example.com/socket",
+      appUrl: "https://app.example.com",
+    });
+  });
+
+  it("rejects invalid JSON", () => {
+    expect(() => parseRuntimeConfig("{")).toThrow(/Invalid desktop runtime config JSON/);
+  });
+
+  it("rejects unsupported schema versions", () => {
+    expect(() =>
+      parseRuntimeConfig(JSON.stringify({ schemaVersion: 2, apiUrl: "https://api.example.com" })),
+    ).toThrow(/schemaVersion/);
+  });
+
+  it("rejects non-http api schemes", () => {
+    expect(() =>
+      parseRuntimeConfig(JSON.stringify({ schemaVersion: 1, apiUrl: "file:///tmp/multica" })),
+    ).toThrow(/apiUrl must use http or https/);
+  });
+
+  it("rejects non-ws websocket schemes", () => {
+    expect(() =>
+      parseRuntimeConfig(
+        JSON.stringify({
+          schemaVersion: 1,
+          apiUrl: "https://api.example.com",
+          wsUrl: "https://api.example.com/ws",
+        }),
+      ),
+    ).toThrow(/wsUrl must use ws or wss/);
+  });
+
+  it("preserves electron-vite dev env precedence", () => {
+    expect(
+      runtimeConfigFromDevEnv({
+        apiUrl: "http://dev-api.example.test:8080/",
+        wsUrl: "ws://dev-api.example.test:8080/ws/",
+        appUrl: "http://dev-app.example.test:3000/",
+      }),
+    ).toEqual({
+      schemaVersion: 1,
+      apiUrl: "http://dev-api.example.test:8080",
+      wsUrl: "ws://dev-api.example.test:8080/ws",
+      appUrl: "http://dev-app.example.test:3000",
+    });
+  });
+
+  it("falls back to local web URL when dev apiUrl is localhost", () => {
+    expect(runtimeConfigFromDevEnv({ apiUrl: "http://localhost:8080" })).toEqual({
+      schemaVersion: 1,
+      apiUrl: "http://localhost:8080",
+      wsUrl: "ws://localhost:8080/ws",
+      appUrl: "http://localhost:3000",
+    });
+  });
+
+  it("derives dev appUrl by stripping the leading api. label", () => {
+    // When the dev renderer is pointed at a remote backend (e.g. a test
+    // environment), copy-link / share URLs must reflect that environment's
+    // public web host, not the api host. Multica's convention exposes the
+    // api at `api.<web-host>`, so stripping the leading label gives the
+    // right web origin without a separate VITE_APP_URL.
+    expect(
+      runtimeConfigFromDevEnv({ apiUrl: "https://api.test.multica.ai" }),
+    ).toEqual({
+      schemaVersion: 1,
+      apiUrl: "https://api.test.multica.ai",
+      wsUrl: "wss://api.test.multica.ai/ws",
+      appUrl: "https://test.multica.ai",
+    });
+  });
+
+  it("dev VITE_APP_URL still wins over apiUrl-derived value", () => {
+    expect(
+      runtimeConfigFromDevEnv({
+        apiUrl: "https://api.test.multica.ai",
+        appUrl: "https://staging.multica.ai",
+      }),
+    ).toEqual({
+      schemaVersion: 1,
+      apiUrl: "https://api.test.multica.ai",
+      wsUrl: "wss://api.test.multica.ai/ws",
+      appUrl: "https://staging.multica.ai",
+    });
+  });
+});

apps/desktop/src/shared/runtime-config.ts

@@ -0,0 +1,179 @@
+export interface RuntimeConfig {
+  schemaVersion: 1;
+  apiUrl: string;
+  wsUrl: string;
+  appUrl: string;
+}
+
+export interface RuntimeConfigError {
+  message: string;
+}
+
+export type RuntimeConfigResult =
+  | { ok: true; config: RuntimeConfig }
+  | { ok: false; error: RuntimeConfigError };
+
+export const DEFAULT_RUNTIME_CONFIG: RuntimeConfig = Object.freeze({
+  schemaVersion: 1,
+  apiUrl: "https://api.multica.ai",
+  wsUrl: "wss://api.multica.ai/ws",
+  appUrl: "https://multica.ai",
+});
+
+const LOCAL_DEV_RUNTIME_CONFIG: RuntimeConfig = Object.freeze({
+  schemaVersion: 1,
+  apiUrl: "http://localhost:8080",
+  wsUrl: "ws://localhost:8080/ws",
+  appUrl: "http://localhost:3000",
+});
+
+export interface RuntimeConfigEnv {
+  apiUrl?: string;
+  wsUrl?: string;
+  appUrl?: string;
+}
+
+export function runtimeConfigFromDevEnv(env: RuntimeConfigEnv): RuntimeConfig {
+  const apiUrl = normalizeHttpUrl(
+    env.apiUrl || LOCAL_DEV_RUNTIME_CONFIG.apiUrl,
+    "VITE_API_URL",
+  );
+  return {
+    schemaVersion: 1,
+    apiUrl,
+    wsUrl: env.wsUrl
+      ? normalizeWsUrl(env.wsUrl, "VITE_WS_URL")
+      : deriveWsUrl(apiUrl),
+    appUrl: env.appUrl
+      ? normalizeHttpUrl(env.appUrl, "VITE_APP_URL")
+      : deriveDevAppUrl(apiUrl),
+  };
+}
+
+export function parseRuntimeConfig(raw: string): RuntimeConfig {
+  let parsed: unknown;
+  try {
+    parsed = JSON.parse(raw);
+  } catch (err) {
+    throw new Error(
+      `Invalid desktop runtime config JSON: ${err instanceof Error ? err.message : "parse failed"}`,
+    );
+  }
+
+  if (!parsed || typeof parsed !== "object" || Array.isArray(parsed)) {
+    throw new Error("Invalid desktop runtime config: expected a JSON object");
+  }
+
+  const obj = parsed as Record<string, unknown>;
+  if (obj.schemaVersion !== 1) {
+    throw new Error("Unsupported desktop runtime config schemaVersion: expected 1");
+  }
+
+  const apiUrl = requiredString(obj.apiUrl, "apiUrl");
+  const appUrl = optionalString(obj.appUrl, "appUrl");
+  const wsUrl = optionalString(obj.wsUrl, "wsUrl");
+
+  const normalizedApiUrl = normalizeHttpUrl(apiUrl, "apiUrl");
+  return {
+    schemaVersion: 1,
+    apiUrl: normalizedApiUrl,
+    wsUrl: wsUrl ? normalizeWsUrl(wsUrl, "wsUrl") : deriveWsUrl(normalizedApiUrl),
+    appUrl: appUrl ? normalizeHttpUrl(appUrl, "appUrl") : deriveAppUrl(normalizedApiUrl),
+  };
+}
+
+export function deriveWsUrl(apiUrl: string): string {
+  const url = new URL(apiUrl);
+  if (url.protocol === "https:") url.protocol = "wss:";
+  else if (url.protocol === "http:") url.protocol = "ws:";
+  else throw new Error("apiUrl must use http or https");
+  url.pathname = joinPath(url.pathname, "/ws");
+  url.search = "";
+  url.hash = "";
+  return trimTrailingSlash(url.toString());
+}
+
+// Convention: api hosts are exposed at `api.<web-host>` (api.multica.ai →
+// multica.ai, api.test.multica.ai → test.multica.ai). Strip the leading
+// `api.` label so a single `apiUrl` configuration produces the right
+// shareable web URL. Hosts that don't match the convention (no leading
+// `api.` label, or short two-label hosts like `api.local`) fall through
+// untouched — those deployments must set `appUrl` explicitly.
+export function deriveAppUrl(apiUrl: string): string {
+  const url = new URL(apiUrl);
+  url.pathname = "";
+  url.search = "";
+  url.hash = "";
+  if (url.hostname.startsWith("api.") && url.hostname.split(".").length >= 3) {
+    url.hostname = url.hostname.slice("api.".length);
+  }
+  return trimTrailingSlash(url.toString());
+}
+
+// Dev variant: when the api host is the local backend (`localhost:8080` /
+// `127.0.0.1:8080`), the renderer is served from a different port (3000),
+// so deriving by host alone is wrong. Fall back to the local dev web URL
+// in that case; for any non-local host (e.g. a remote test environment),
+// trust the production-style derivation so `apiUrl=https://api.test.x`
+// yields `appUrl=https://test.x` without a separate VITE_APP_URL.
+export function deriveDevAppUrl(apiUrl: string): string {
+  const url = new URL(apiUrl);
+  if (url.hostname === "localhost" || url.hostname === "127.0.0.1") {
+    return LOCAL_DEV_RUNTIME_CONFIG.appUrl;
+  }
+  return deriveAppUrl(apiUrl);
+}
+
+function requiredString(value: unknown, field: string): string {
+  if (typeof value !== "string" || value.trim().length === 0) {
+    throw new Error(`Invalid desktop runtime config: ${field} must be a non-empty string`);
+  }
+  return value;
+}
+
+function optionalString(value: unknown, field: string): string | undefined {
+  if (value === undefined) return undefined;
+  if (typeof value !== "string" || value.trim().length === 0) {
+    throw new Error(`Invalid desktop runtime config: ${field} must be a non-empty string when set`);
+  }
+  return value;
+}
+
+function normalizeHttpUrl(value: string, field: string): string {
+  let url: URL;
+  try {
+    url = new URL(value.trim());
+  } catch {
+    throw new Error(`Invalid desktop runtime config: ${field} must be a valid URL`);
+  }
+  if (url.protocol !== "http:" && url.protocol !== "https:") {
+    throw new Error(`Invalid desktop runtime config: ${field} must use http or https`);
+  }
+  url.search = "";
+  url.hash = "";
+  return trimTrailingSlash(url.toString());
+}
+
+function normalizeWsUrl(value: string, field: string): string {
+  let url: URL;
+  try {
+    url = new URL(value.trim());
+  } catch {
+    throw new Error(`Invalid desktop runtime config: ${field} must be a valid URL`);
+  }
+  if (url.protocol !== "ws:" && url.protocol !== "wss:") {
+    throw new Error(`Invalid desktop runtime config: ${field} must use ws or wss`);
+  }
+  url.search = "";
+  url.hash = "";
+  return trimTrailingSlash(url.toString());
+}
+
+function joinPath(base: string, suffix: string): string {
+  const normalizedBase = base.endsWith("/") ? base.slice(0, -1) : base;
+  return `${normalizedBase}${suffix}`;
+}
+
+function trimTrailingSlash(value: string): string {
+  return value.replace(/\/+$/, "");
+}

apps/desktop/test/setup.ts

@@ -1 +1,38 @@
 import "@testing-library/jest-dom/vitest";
+
+function createMemoryStorage(): Storage {
+  const values = new Map<string, string>();
+
+  return {
+    get length() {
+      return values.size;
+    },
+    clear: () => values.clear(),
+    getItem: (key: string) => values.get(key) ?? null,
+    key: (index: number) => Array.from(values.keys())[index] ?? null,
+    removeItem: (key: string) => {
+      values.delete(key);
+    },
+    setItem: (key: string, value: string) => {
+      values.set(key, value);
+    },
+  };
+}
+
+const localStorageIsUsable =
+  typeof globalThis.localStorage?.getItem === "function" &&
+  typeof globalThis.localStorage?.setItem === "function" &&
+  typeof globalThis.localStorage?.removeItem === "function" &&
+  typeof globalThis.localStorage?.clear === "function";
+
+if (!localStorageIsUsable) {
+  const storage = createMemoryStorage();
+  Object.defineProperty(globalThis, "localStorage", {
+    configurable: true,
+    value: storage,
+  });
+  Object.defineProperty(window, "localStorage", {
+    configurable: true,
+    value: storage,
+  });
+}

apps/desktop/vitest.config.ts

@@ -1,8 +1,14 @@
+import { resolve } from "path";
 import { defineConfig } from "vitest/config";
 import react from "@vitejs/plugin-react";
 
 export default defineConfig({
   plugins: [react()],
+  resolve: {
+    alias: {
+      "@": resolve(__dirname, "src/renderer/src"),
+    },
+  },
   test: {
     globals: true,
     include: ["src/**/*.test.{ts,tsx}", "scripts/**/*.test.mjs"],

apps/docs/app/[lang]/[...slug]/page.tsx

@@ -9,6 +9,14 @@ import { notFound } from "next/navigation";
 import defaultMdxComponents from "fumadocs-ui/mdx";
 import type { Metadata } from "next";
 import { docsAlternates } from "@/lib/site";
+import { i18n, type Lang } from "@/lib/i18n";
+import { DocsLocaleProvider, LocaleLink } from "@/components/locale-link";
+
+function asLang(lang: string): Lang {
+  return (i18n.languages as readonly string[]).includes(lang)
+    ? (lang as Lang)
+    : (i18n.defaultLanguage as Lang);
+}
 
 export default async function Page(props: {
   params: Promise<{ lang: string; slug: string[] }>;
@@ -18,13 +26,16 @@ export default async function Page(props: {
   if (!page) notFound();
 
   const MDX = page.data.body;
+  const lang = asLang(params.lang);
 
   return (
     <DocsPage toc={page.data.toc}>
       <DocsTitle>{page.data.title}</DocsTitle>
       <DocsDescription>{page.data.description}</DocsDescription>
       <DocsBody>
-        <MDX components={{ ...defaultMdxComponents }} />
+        <DocsLocaleProvider lang={lang}>
+          <MDX components={{ ...defaultMdxComponents, a: LocaleLink }} />
+        </DocsLocaleProvider>
       </DocsBody>
     </DocsPage>
   );

apps/docs/app/[lang]/page.tsx

@@ -8,6 +8,7 @@ import { Byline, NumberedCards, NumberedCard, NumberedSteps, Step } from "@/comp
 import { i18n, type Lang } from "@/lib/i18n";
 import { homeCopy } from "@/lib/translations";
 import { docsAlternates } from "@/lib/site";
+import { DocsLocaleProvider, LocaleLink } from "@/components/locale-link";
 
 function asLang(lang: string): Lang {
   return (i18n.languages as readonly string[]).includes(lang)
@@ -52,15 +53,18 @@ export default async function Page({
       />
       <Byline items={[...copy.byline]} />
       <DocsBody>
-        <MDX
-          components={{
-            ...defaultMdxComponents,
-            NumberedCards,
-            NumberedCard,
-            NumberedSteps,
-            Step,
-          }}
-        />
+        <DocsLocaleProvider lang={lang}>
+          <MDX
+            components={{
+              ...defaultMdxComponents,
+              a: LocaleLink,
+              NumberedCards,
+              NumberedCard,
+              NumberedSteps,
+              Step,
+            }}
+          />
+        </DocsLocaleProvider>
       </DocsBody>
     </DocsPage>
   );

apps/docs/components/editorial.tsx

@@ -1,5 +1,9 @@
+"use client";
+
 import Link from "next/link";
 import type { ReactNode } from "react";
+import { useDocsLocale } from "@/components/locale-link";
+import { prefixLocale } from "@/lib/locale-link";
 
 /**
  * Byline — editorial metadata strip with ruled top + bottom borders.
@@ -55,9 +59,10 @@ export function NumberedCard({
   tag?: string;
   children: ReactNode;
 }) {
+  const lang = useDocsLocale();
   return (
     <Link
-      href={href}
+      href={prefixLocale(href, lang)}
       className="group flex flex-col gap-2.5 border-r border-border px-0 py-5 pr-4 no-underline last:border-r-0 md:px-4 md:first:pl-0 md:last:pr-0"
     >
       <div className="font-mono text-[0.6875rem] uppercase tracking-[0.08em] text-muted-foreground">

apps/docs/components/locale-link.tsx

@@ -0,0 +1,48 @@
+"use client";
+
+import Link from "next/link";
+import {
+  createContext,
+  useContext,
+  type AnchorHTMLAttributes,
+  type ReactNode,
+} from "react";
+import { i18n, type Lang } from "@/lib/i18n";
+import { prefixLocale } from "@/lib/locale-link";
+
+const DocsLocaleContext = createContext<Lang>(i18n.defaultLanguage as Lang);
+
+// Wraps the rendered MDX subtree so descendant <LocaleLink>s and any
+// editorial component using `useDocsLocale()` know which language the page
+// was rendered in. Mounted at each docs page entry; never elsewhere.
+export function DocsLocaleProvider({
+  lang,
+  children,
+}: {
+  lang: Lang;
+  children: ReactNode;
+}) {
+  return (
+    <DocsLocaleContext.Provider value={lang}>
+      {children}
+    </DocsLocaleContext.Provider>
+  );
+}
+
+export function useDocsLocale(): Lang {
+  return useContext(DocsLocaleContext);
+}
+
+// Drop-in replacement for the MDX-rendered `<a>` element. Keeps the same
+// surface shape as the default `a` from `defaultMdxComponents` but routes
+// internal links through the locale prefixer + next/link so client-side
+// navigation stays inside the active locale.
+export function LocaleLink({
+  href,
+  ...rest
+}: AnchorHTMLAttributes<HTMLAnchorElement> & { href?: string }) {
+  const lang = useDocsLocale();
+  if (!href) return <a {...rest} />;
+  const final = prefixLocale(href, lang);
+  return <Link href={final} {...rest} />;
+}

apps/docs/content/docs/assigning-issues.mdx

@@ -32,9 +32,10 @@ The command-line equivalent:
 
 ```bash
 multica issue assign MUL-42 --to alice
+multica issue assign MUL-42 --to-id 5fb87ac7-23b5-4a7a-81fa-ed295a54545d
 ```
 
-`--to` takes a member username or an agent name. Giving agents memorable names makes this step smoother — if multiple agents share a name in the workspace, the first one listed wins, so rename before assigning.
+`--to` takes a member username or an agent name (fuzzy match). When names overlap — e.g. an agent `J` alongside `Cursor - J` — pass `--to-id <uuid>` instead, using the `user_id` (member) or `id` (agent) from `multica workspace members --output json` / `multica agent list --output json`. UUID matching is strict and unambiguous, which is what you want from scripts and from agents driving the CLI. `--to` and `--to-id` are mutually exclusive.
 
 Unassign:
 

apps/docs/content/docs/assigning-issues.zh.mdx

@@ -32,9 +32,10 @@ import { Callout } from "fumadocs-ui/components/callout";
 
 ```bash
 multica issue assign MUL-42 --to alice
+multica issue assign MUL-42 --to-id 5fb87ac7-23b5-4a7a-81fa-ed295a54545d
 ```
 
-`--to` 后跟成员用户名或智能体名字。给智能体起个好记的名字会让这一步顺很多——工作区里重名的会按列出顺序选第一个，建议先改名再分配。
+`--to` 后跟成员用户名或智能体名字（模糊匹配）。如果工作区里有同名 / 互相含子串的成员或智能体（例如 agent `J` 旁边还有 `Cursor - J`），改用 `--to-id <uuid>`：UUID 来自 `multica workspace members --output json` 的 `user_id` 或 `multica agent list --output json` 的 `id`，是唯一精确的方式，特别适合脚本和驱动 CLI 的智能体。`--to` 和 `--to-id` 互斥。
 
 取消分配：
 

apps/docs/content/docs/cli.mdx

@@ -40,20 +40,25 @@ For the difference between token types, see [Authentication and tokens](/auth-to
 | `multica workspace list` | List every workspace you can access |
 | `multica workspace get <slug>` | Show details for one workspace |
 | `multica workspace members` | List members of the current workspace |
+| `multica workspace update <id> --name "..." [--description "..."] [--context "..."] [--issue-prefix "..."]` | Update workspace metadata (admin/owner). Long fields accept `--description-stdin` / `--context-stdin`. |
 
 ## Issues and projects
 
+<Callout type="info">
+`list` commands (`multica issue list`, `autopilot list`, `project list`, etc.) print short, copy-paste-ready IDs by default — issue keys like `MUL-123` for issues, short UUID prefixes for the rest. The `<id>` argument on the follow-up commands below accepts either the short ID or the full UUID, so the typical flow is `multica issue list` → copy the key → `multica issue get MUL-123`. Pass `--full-id` to a list command when you need the canonical UUID.
+</Callout>
+
 | Command | Purpose |
 |---|---|
-| `multica issue list` | List issues |
-| `multica issue get <id>` | Show a single issue |
+| `multica issue list` | List issues (prints copy-paste-ready issue keys) |
+| `multica issue get <id>` | Show a single issue (accepts an issue key or a UUID) |
 | `multica issue create --title "..."` | Create a new issue |
 | `multica issue update <id> ...` | Update an issue (status, priority, assignee, etc.) |
 | `multica issue assign <id> --agent <slug>` | Assign to an agent (triggers a task immediately) |
 | `multica issue status <id> --set <status>` | Shortcut to change status |
 | `multica issue search <query>` | Keyword search |
 | `multica issue runs <id>` | Show agent runs on an issue |
-| `multica issue rerun <id>` | Rerun the most recent agent task |
+| `multica issue rerun <id>` | Re-enqueue a fresh task for the issue's current agent assignee |
 | `multica issue comment <id> ...` | Nested: view / post comments |
 | `multica issue subscriber <id> ...` | Nested: subscribe / unsubscribe |
 | `multica project list/get/create/update/delete/status` | Project CRUD |
@@ -98,7 +103,6 @@ For the difference between token types, see [Authentication and tokens](/auth-to
 | `multica runtime list` | List runtimes in the current workspace |
 | `multica runtime usage` | Show resource usage |
 | `multica runtime activity` | Recent activity log |
-| `multica runtime ping <id>` | Ping a runtime to check it's online |
 | `multica runtime update <id> ...` | Update a runtime's configuration |
 
 ## Miscellaneous

apps/docs/content/docs/cli.zh.mdx

@@ -40,20 +40,25 @@ Token 类型的详细区分见 [认证与令牌](/auth-tokens)。
 | `multica workspace list` | 列出你有权访问的所有工作区 |
 | `multica workspace get <slug>` | 查看一个工作区的详情 |
 | `multica workspace members` | 列出当前工作区的成员 |
+| `multica workspace update <id> --name "..." [--description "..."] [--context "..."] [--issue-prefix "..."]` | 修改 workspace 元数据（admin/owner 权限）。长文本可用 `--description-stdin` / `--context-stdin`。 |
 
 ## Issue 和 Project
 
+<Callout type="info">
+`list` 类命令（`multica issue list`、`autopilot list`、`project list` 等）表格里默认显示**可直接复制**的短 ID：issue 是 key（如 `MUL-123`），其余资源是 UUID 短前缀。下面表格里的 `<id>` 同时接受短 ID 和完整 UUID，所以典型用法是 `multica issue list` → 复制 key → `multica issue get MUL-123`。需要完整 UUID 时给 `list` 加 `--full-id`。
+</Callout>
+
 | 命令 | 用途 |
 |---|---|
-| `multica issue list` | 列出 issue |
-| `multica issue get <id>` | 查看单条 issue |
+| `multica issue list` | 列出 issue（默认显示可复制的 issue key） |
+| `multica issue get <id>` | 查看单条 issue（接受 issue key 或 UUID） |
 | `multica issue create --title "..."` | 创建新 issue |
 | `multica issue update <id> ...` | 修改 issue（状态、优先级、分配人等） |
 | `multica issue assign <id> --agent <slug>` | 分配给智能体（立即触发任务） |
 | `multica issue status <id> --set <status>` | 快捷改状态 |
 | `multica issue search <query>` | 关键字搜索 |
 | `multica issue runs <id>` | 查看 issue 上智能体跑过的任务 |
-| `multica issue rerun <id>` | 重跑最近一次智能体任务 |
+| `multica issue rerun <id>` | 给该 issue 当前的智能体分配人重新创建一条任务 |
 | `multica issue comment <id> ...` | 嵌套：看 / 发评论 |
 | `multica issue subscriber <id> ...` | 嵌套：订阅 / 取消订阅 |
 | `multica project list/get/create/update/delete/status` | Project CRUD |
@@ -98,7 +103,6 @@ Token 类型的详细区分见 [认证与令牌](/auth-tokens)。
 | `multica runtime list` | 列出当前工作区的 runtime |
 | `multica runtime usage` | 查看资源使用情况 |
 | `multica runtime activity` | 近期活动记录 |
-| `multica runtime ping <id>` | 立即戳一次 runtime 检查在线 |
 | `multica runtime update <id> ...` | 更新 runtime 配置 |
 
 ## 杂项

apps/docs/content/docs/cli/reference.zh.mdx

@@ -18,10 +18,10 @@ Opens your browser for OAuth authentication, creates a 90-day personal access to
 ### Token Login
 
 ```bash
-multica login --token
+multica login --token <mul_...>
 ```
 
-Authenticate by pasting a personal access token directly. Useful for headless environments.
+Authenticate using a personal access token directly. Useful for headless environments. Pass `--token=` with an empty value to be prompted interactively (so the token never lands in shell history).
 
 ### Check Status
 
@@ -213,6 +213,28 @@ multica workspace get <workspace-id> --output json
 multica workspace members <workspace-id>
 ```
 
+### Update Workspace
+
+需要 admin 或 owner 权限。所有字段都是部分更新（PATCH 语义）：未传的字段保持不变。
+
+```bash
+multica workspace update <workspace-id> --name "Acme Eng"
+multica workspace update <workspace-id> \
+  --description "Engineering team workspace" \
+  --issue-prefix ENG
+```
+
+长文本走 stdin（保留换行/反斜杠）：
+
+```bash
+cat <<'CTX' | multica workspace update <workspace-id> --context-stdin
+我们是一支 5 人 AI-native 团队。
+工作语言：中文 + 英文混合。
+CTX
+```
+
+可编辑字段：`--name`、`--description` / `--description-stdin`、`--context` / `--context-stdin`、`--issue-prefix`。`slug` 创建后只读，不暴露在 CLI。`--description` 与 `--description-stdin`（以及 `context` 同名对）互斥。未传任何字段 flag 时命令拒绝执行，避免空 PATCH 触发无意义的 workspace 更新事件。`--issue-prefix ""` 也会被拒绝：当前后端在 prefix 为空时静默跳过该字段，CLI 在本地拦下避免“看似成功的 no-op”。
+
 ## Issues
 
 ### List Issues
@@ -221,25 +243,31 @@ multica workspace members <workspace-id>
 multica issue list
 multica issue list --status in_progress
 multica issue list --priority urgent --assignee "Agent Name"
+multica issue list --assignee-id 5fb87ac7-23b5-4a7a-81fa-ed295a54545d
+multica issue list --full-id
 multica issue list --limit 20 --output json
 ```
 
-Available filters: `--status`, `--priority`, `--assignee`, `--project`, `--limit`.
+表格输出默认显示可直接复制到后续命令的 issue `KEY`（例如 `MUL-123`）；需要完整 UUID 时使用 `--full-id`。Available filters: `--status`, `--priority`, `--assignee` / `--assignee-id`, `--project`, `--limit`. 在重名 workspace 下用 `--assignee-id <uuid>` 可以精确锁定一个成员或 agent。
 
 ### Get Issue
 
 ```bash
-multica issue get <id>
+multica issue get MUL-123
+multica issue get <uuid>
 multica issue get <id> --output json
 ```
 
+`<id>` 同时接受 issue key（`multica issue list` 表格里直接显示，例如 `MUL-123`）和完整 UUID（给 `list` 加 `--full-id` 可显示）。同样的规则适用于下面 `update` / `assign` / `status` / `comment` / `subscriber` / `runs` 等接受 `<id>` 的命令。
+
 ### Create Issue
 
 ```bash
 multica issue create --title "Fix login bug" --description "..." --priority high --assignee "Lambda"
+multica issue create --title "Fix login bug" --assignee-id 5fb87ac7-23b5-4a7a-81fa-ed295a54545d
 ```
 
-Flags: `--title` (required), `--description`, `--status`, `--priority`, `--assignee`, `--parent`, `--project`, `--due-date`.
+Flags: `--title` (required), `--description`, `--status`, `--priority`, `--assignee` / `--assignee-id`, `--parent`, `--project`, `--due-date`. 脚本里如果已经拿到了 UUID（例如来自 `multica workspace members --output json`），传 `--assignee-id <uuid>`（与 `--assignee` 互斥）以精确锁定。
 
 ### Update Issue
 
@@ -251,9 +279,12 @@ multica issue update <id> --title "New title" --priority urgent
 
 ```bash
 multica issue assign <id> --to "Lambda"
+multica issue assign <id> --to-id 5fb87ac7-23b5-4a7a-81fa-ed295a54545d
 multica issue assign <id> --unassign
 ```
 
+`--to-id <uuid>`（与 `--to` 互斥）按 UUID 精确分配；适合重名 workspace 下脚本化场景。
+
 ### Change Status
 
 ```bash
@@ -283,16 +314,20 @@ multica issue comment delete <comment-id>
 ```bash
 # List all execution runs for an issue
 multica issue runs <issue-id>
+multica issue runs <issue-id> --full-id
 multica issue runs <issue-id> --output json
 
 # View messages for a specific execution run
 multica issue run-messages <task-id>
+multica issue run-messages <short-task-id> --issue <issue-id>
 multica issue run-messages <task-id> --output json
 
 # Incremental fetch (only messages after a given sequence number)
 multica issue run-messages <task-id> --since 42 --output json
 ```
 
+`runs` 的表格输出默认显示 task UUID 短前缀；需要完整 task UUID 时使用 `--full-id`。`run-messages` 可直接接受完整 task UUID；从 `runs` 表格复制短前缀时需要同时传 `--issue <issue-id>`，CLI 只会在该 issue 的 runs 内解析。
+
 ## Projects
 
 Projects group related issues (e.g. a sprint, an epic, a workstream). Every project

apps/docs/content/docs/cloud-quickstart.mdx

@@ -99,7 +99,7 @@ Assign the issue to the agent you just created — click its avatar in the web U
 multica issue assign MUL-1 --to my-agent-name
 ```
 
-`--to` takes the **name** of an agent or member. A substring match works — if the agent is called `my-code-reviewer`, `reviewer` resolves to it.
+`--to` takes the **name** of an agent or member. A substring match works — if the agent is called `my-code-reviewer`, `reviewer` resolves to it. If your workspace has overlapping names, pass `--to-id <uuid>` instead (mutually exclusive with `--to`); look up the UUID via `multica agent list --output json` or `multica workspace members --output json`.
 
 **What happens next from the daemon**:
 

apps/docs/content/docs/cloud-quickstart.zh.mdx

@@ -99,7 +99,7 @@ multica issue create --title "给 README 加一段 ASCII 架构图"
 multica issue assign MUL-1 --to my-agent-name
 ```
 
-`--to` 后面填智能体或成员的**名字**，子串就行——如果智能体叫 `my-code-reviewer`，填 `reviewer` 也能命中。
+`--to` 后面填智能体或成员的**名字**，子串就行——如果智能体叫 `my-code-reviewer`，填 `reviewer` 也能命中。如果工作区里名字相互重叠或冲突，改用 `--to-id <uuid>`（与 `--to` 互斥）；UUID 来自 `multica agent list --output json` 或 `multica workspace members --output json`。
 
 **接下来守护进程会**：
 

apps/docs/content/docs/daemon-runtimes.mdx

@@ -37,7 +37,7 @@ Common commands:
 
 Full CLI reference in [CLI commands](/cli).
 
-**The desktop app ships with a daemon.** If you use the [desktop app](/desktop-app), you don't need to run `multica daemon start` manually — it launches the daemon automatically on startup.
+**The desktop app ships with a daemon.** If you use the [desktop app](/desktop-app), you don't need to run `multica daemon start` manually — it launches the daemon automatically on startup. See the [Desktop app](/desktop-app) page for which option fits your workflow.
 
 ## Why one machine has multiple runtimes
 

apps/docs/content/docs/daemon-runtimes.zh.mdx

@@ -37,7 +37,7 @@ multica daemon start
 
 完整 CLI 参考见 [CLI 命令速查](/cli)。
 
-**桌面应用自带守护进程。**用 [桌面应用](/desktop-app) 就不必手动 `multica daemon start`——它启动时会自动拉起守护进程。
+**桌面应用自带守护进程。**用 [桌面应用](/desktop-app) 就不必手动 `multica daemon start`——它启动时会自动拉起守护进程。哪种方式更适合你的工作流，详见 [桌面应用](/desktop-app) 页面。
 
 ## 为什么一台机器会有多个运行时
 

apps/docs/content/docs/desktop-app.mdx

@@ -5,7 +5,7 @@ description: What Multica Desktop is, how it differs from the web app, and when
 
 import { Callout } from "fumadocs-ui/components/callout";
 
-Multica Desktop is a native desktop app for macOS, Windows, and Linux. It talks to the same backend as the web app and shows the same data, but it adds a few things the browser can't: **independent tab groups per [workspace](/workspaces)**, **automatic [daemon](/daemon-runtimes) startup**, and **one-click upgrades**.
+Multica Desktop is a native desktop app for macOS, Windows, and Linux. For the environment it is configured for, it talks to the same backend as the web app and shows the same data. By default Desktop uses Multica Cloud; self-hosted instances can be configured with a local runtime config file. Desktop also adds a few things the browser can't: **independent tab groups per [workspace](/workspaces)**, **automatic [daemon](/daemon-runtimes) startup**, and **one-click upgrades**.
 
 ## Desktop or web — which to pick
 
@@ -66,25 +66,34 @@ Grab the installer for your platform from the [Multica downloads page](https://m
 
 On first launch you'll need to sign in — the same email + verification code flow as the web app. Once you're in, Desktop syncs your workspace list automatically.
 
-<Callout type="warning">
-**Released Desktop builds are pinned to Multica Cloud.** The backend, websocket, and web URLs are baked in at build time (`VITE_API_URL` / `VITE_WS_URL` / `VITE_APP_URL`) — there is no in-app option to point Desktop at a self-hosted instance. To use Desktop against a self-hosted backend you need to build it yourself:
+<Callout type="info">
+**Desktop defaults to Multica Cloud, but can be pointed at a self-hosted instance with a local config file.** There is still no in-app "connect to self-host" picker. Desktop reads `~/.multica/desktop.json` before the renderer starts; if the file is missing, it uses the Cloud defaults.
 
-```bash
-git clone https://github.com/multica-ai/multica.git
-cd multica
-# Edit apps/desktop/.env.production:
-#   VITE_API_URL=https://api.your-domain
-#   VITE_WS_URL=wss://api.your-domain/ws
-#   VITE_APP_URL=https://your-domain
-pnpm install
-pnpm --filter @multica/desktop package
+Minimal self-host config:
+
+```json
+{
+  "schemaVersion": 1,
+  "apiUrl": "https://api.your-domain"
+}
 ```
 
-If you'd rather not build from source, the supported self-hosted path is **web frontend + CLI** — see [Self-host quickstart](/self-host-quickstart). Runtime backend configuration in Desktop is tracked in [issue #1371](https://github.com/multica-ai/multica/issues/1371).
+`apiUrl` is required and must use `http` or `https`. Desktop derives `wsUrl` as `/ws` on the same origin (`wss` for `https`, `ws` for `http`) and derives `appUrl` from the API origin. If your deployment uses different origins, set them explicitly:
+
+```json
+{
+  "schemaVersion": 1,
+  "apiUrl": "https://api.your-domain",
+  "wsUrl": "wss://api.your-domain/ws",
+  "appUrl": "https://your-domain"
+}
+```
+
+If `desktop.json` exists but is invalid, Desktop fails closed and shows a blocking config error instead of silently falling back to Cloud. For development builds, `VITE_API_URL` / `VITE_WS_URL` / `VITE_APP_URL` still take precedence during `electron-vite dev`. Runtime Desktop self-host configuration was implemented for [issue #1371](https://github.com/multica-ai/multica/issues/1371).
 </Callout>
 
 ## Next steps
 
 - [Cloud Quickstart](/cloud-quickstart) — the Cloud onboarding flow for Desktop
-- [Self-Host Quickstart](/self-host-quickstart) — running your own backend (Desktop against self-host requires a custom build, see the callout above)
+- [Self-Host Quickstart](/self-host-quickstart) — running your own backend and connecting with the CLI or Desktop runtime config
 - [Daemon and runtimes](/daemon-runtimes) — how the daemon works (Desktop starts it for you, but the behavior is the same)

apps/docs/content/docs/desktop-app.zh.mdx

@@ -5,7 +5,7 @@ description: Multica Desktop 是什么、和 Web 有什么区别、什么时候
 
 import { Callout } from "fumadocs-ui/components/callout";
 
-Multica Desktop 是原生桌面应用——macOS / Windows / Linux 三个平台。它和 Web 版连同一个后端，看到的数据完全一样，但给了几个 Web 做不到的能力：**[工作区](/workspaces) 独立的多标签页**、**自动启动 [守护进程](/daemon-runtimes)**、**一键升级**。
+Multica Desktop 是原生桌面应用——macOS / Windows / Linux 三个平台。对它当前配置的环境来说，它和 Web 版连同一个后端、看到的数据完全一样。Desktop 默认使用 Multica Cloud；自部署实例可以通过本地运行时配置文件接入。它还给了几个 Web 做不到的能力：**[工作区](/workspaces) 独立的多标签页**、**自动启动 [守护进程](/daemon-runtimes)**、**一键升级**。
 
 ## Desktop 和 Web 该用哪个
 
@@ -66,25 +66,34 @@ macOS 版本已经签名 + 公证，第一次打开不会有"未知开发者"的
 
 安装后第一次打开需要登录——和 Web 版一样的 email + 验证码流程。登录成功后 Desktop 自动把工作区列表同步下来。
 
-<Callout type="warning">
-**发布版的 Desktop 是锁死连 Multica Cloud 的**。后端 / WebSocket / Web 前端 URL（`VITE_API_URL` / `VITE_WS_URL` / `VITE_APP_URL`）在构建时就写死了，应用内**没有切换后端的入口**。要让 Desktop 连自部署后端，需要你自己从源码 build：
+<Callout type="info">
+**Desktop 默认连接 Multica Cloud，但可以通过本地配置文件指向自部署实例。** 应用内仍然没有“连接自部署”的切换入口。Desktop 会在 renderer 启动前读取 `~/.multica/desktop.json`；如果这个文件不存在，就使用 Cloud 默认值。
 
-```bash
-git clone https://github.com/multica-ai/multica.git
-cd multica
-# 编辑 apps/desktop/.env.production：
-#   VITE_API_URL=https://api.your-domain
-#   VITE_WS_URL=wss://api.your-domain/ws
-#   VITE_APP_URL=https://your-domain
-pnpm install
-pnpm --filter @multica/desktop package
+最小自部署配置：
+
+```json
+{
+  "schemaVersion": 1,
+  "apiUrl": "https://api.your-domain"
+}
 ```
 
-不想自己 build 的话，自部署的官方路径是 **Web 前端 + CLI**——见 [自部署快速上手](/self-host-quickstart)。Desktop 运行时切换后端的能力跟踪在 [issue #1371](https://github.com/multica-ai/multica/issues/1371)。
+`apiUrl` 是必填项，必须使用 `http` 或 `https`。Desktop 会自动从它推导 `wsUrl`（同源 `/ws`，`https` 对应 `wss`，`http` 对应 `ws`）和 `appUrl`（API 的同源地址）。如果你的部署使用不同域名，可以显式设置：
+
+```json
+{
+  "schemaVersion": 1,
+  "apiUrl": "https://api.your-domain",
+  "wsUrl": "wss://api.your-domain/ws",
+  "appUrl": "https://your-domain"
+}
+```
+
+如果 `desktop.json` 存在但内容无效，Desktop 会 fail closed，显示阻塞式配置错误，而不是悄悄回退到 Cloud。开发构建里，`electron-vite dev` 仍然优先使用 `VITE_API_URL` / `VITE_WS_URL` / `VITE_APP_URL`。Desktop 运行时自部署配置能力对应 [issue #1371](https://github.com/multica-ai/multica/issues/1371)。
 </Callout>
 
 ## 下一步
 
 - [Cloud Quickstart](/cloud-quickstart) —— Desktop 版的 Cloud 接入流程
-- [Self-Host Quickstart](/self-host-quickstart) —— 自部署后端（Desktop 连自部署需要自行构建，见上方提示）
+- [Self-Host Quickstart](/self-host-quickstart) —— 自部署后端，并通过 CLI 或 Desktop 运行时配置连接
 - [守护进程与运行时](/daemon-runtimes) —— 守护进程机制（Desktop 自动起它，但行为一样）

apps/docs/content/docs/developers/conventions.mdx

@@ -0,0 +1,301 @@
+---
+title: Conventions
+description: Single source of truth for code naming, i18n translation glossary, and Chinese voice guide.
+---
+
+This page is the single source of truth for code naming, the i18n translation glossary, and the Chinese voice guide. Anything that used to live in `packages/views/locales/glossary.md` or in scattered comments now lives here.
+
+If you write Multica code, change a translation, or write Chinese product copy, this is the page to reference.
+
+---
+
+## 1. Code naming
+
+### Routes
+
+Pre-workspace routes (the routes that exist before the user is in a workspace) MUST use either a single word or the `/{noun}/{verb}` pattern.
+
+- ✅ `/login`, `/inbox`, `/workspaces/new`
+- ❌ `/new-workspace`, `/create-team`, `/accept-invite`
+
+Hyphenated word groups at the root collide with user-chosen workspace slugs and force endless reserved-slug audits. Reserving the noun (`workspaces`) automatically protects the entire `/workspaces/*` subtree.
+
+### Workspace-scoped routes
+
+Always live under `/{slug}/{section}` — `/{slug}/issues`, `/{slug}/agents`, `/{slug}/settings`. Never duplicate workspace routing logic; use `useNavigation().push()` from shared code, never framework-specific link APIs.
+
+### Packages and modules
+
+The monorepo enforces strict package boundaries:
+
+| Package | May depend on | Must NOT depend on |
+| --- | --- | --- |
+| `packages/core` | nothing app-specific | `react-dom`, `localStorage`, `process.env`, `next/*`, UI libraries |
+| `packages/ui` | nothing | `@multica/core`, business logic |
+| `packages/views` | `core/`, `ui/` | `next/*`, `react-router-dom`, stores |
+| `apps/web/platform/` | `next/*` | other apps |
+| `apps/desktop/.../platform/` | `react-router-dom`, electron | other apps |
+
+If logic appears in both apps, it MUST be extracted to a shared package. There are no exceptions for "small" duplication.
+
+### Files and components
+
+- Files: `kebab-case.tsx` / `kebab-case.ts` (e.g. `agent-row-actions.tsx`)
+- Components: `PascalCase` (e.g. `AgentRowActions`)
+- Hooks: `useCamelCase` (e.g. `useWorkspaceId`)
+- Tests: colocated as `<file>.test.ts(x)`
+- Stores (Zustand): `<feature>-store.ts`, exported as `use<Feature>Store`
+
+### Database (Go + sqlc)
+
+- Tables: `snake_case` singular (`user`, `workspace`, `agent_runtime`)
+- Columns: `snake_case` (`workspace_id`, `created_at`, `last_seen_at`)
+- Foreign keys: `<table>_id`
+- Booleans: `is_<state>` or `<state>_at` (timestamp form preferred for state changes)
+- Migration files: `NNN_descriptive_name.up.sql` + `.down.sql` — always provide both directions
+
+### Go
+
+- Standard `gofmt` + `go vet`. No exceptions.
+- Handler files mirror domain: `agent.go`, `auth.go`, `runtime.go`
+- Tests: `<file>_test.go` colocated
+- For UUID parsing in handlers, follow the rule in the root `CLAUDE.md` — `parseUUIDOrBadRequest` for boundary input, `parseUUID` (panicking) for trusted round-trips, never `util.ParseUUID` directly without checking the error.
+
+### TypeScript
+
+- API responses on the wire are `snake_case`; the api client converts to `camelCase` at the boundary. Inside TS code, **always camelCase**.
+- Types: `PascalCase` (`Issue`, `AgentRuntime`); never `IPrefix`, never `_t` suffix.
+- Enums: prefer string literal unions; reserve `enum` for runtime-iterable cases.
+- TanStack Query keys: factory functions in `<feature>/queries.ts`, e.g. `issueKeys.detail(id)`.
+
+### Issue keys
+
+Every issue has a human-readable key like `MUL-123`: workspace `issue_prefix` (3 letters, uppercase) + sequence number. The prefix is set at workspace creation and is never changed afterward.
+
+### Comments in code
+
+English only. The repo enforces this for both Go and TypeScript. If you find a Chinese comment in code, it's a bug — replace it.
+
+### Commit messages
+
+Conventional format: `feat(scope)`, `fix(scope)`, `refactor(scope)`, `docs`, `test(scope)`, `chore(scope)`. Atomic commits grouped by intent.
+
+---
+
+## 2. i18n translation glossary
+
+This is the **mandatory** glossary for every translation PR. It used to live at `packages/views/locales/glossary.md`; that file is now a stub pointing here.
+
+### The core distinction: entity vs concept
+
+Multica's product nouns split into two categories:
+
+- **Entity** — has a URL, a database row, an API type. In Chinese text, render as **lowercase English** so it visually reads like a type name and signals "this is a Multica system entity".
+- **Concept** — generic noun, not a database entity. **Translate fully** so Chinese users don't see jagged English embedded in flowing text.
+
+This rule is aligned with `apps/docs/content/docs/*.zh.mdx` — the docs are the de facto Chinese voice standard and have been battle-tested across 20+ pages.
+
+### Entities — mixed rule (`issue` / `skill` / `task`)
+
+`issue` / `skill` / `task` are Multica's core entities. They have schema columns, API fields, and product UI labels that are all English. In Chinese text, they follow a **mixed rule** — what to use depends on where the word appears:
+
+| Context | Render | Example |
+| --- | --- | --- |
+| **UI strings, state names, code references** | lowercase English | "排队中的 task"、"创建子 issue"、"为智能体注入 skill" |
+| **Doc titles / section headings** | Title-case English **or** the Chinese term | "Issue 与 project"、"Skills"、"执行任务" |
+| **Long-form doc prose, when the entity is the running subject** | Chinese term, with English in parentheses on first mention | "**执行任务**（task）是智能体每一次工作的单位" |
+| **API / DB fields** | always `task` / `issue` / `skill` | `task_id`, `issue_status`, `skill_uuid` |
+
+Chinese term reference:
+
+- `task` ↔ `执行任务` (or shortened to `任务` once context is clear)
+- `issue` has no settled Chinese translation — leave English; titles may capitalize as `Issue`
+- `skill` has no settled Chinese translation — leave English; titles may capitalize as `Skills`
+
+**Why `issue` / `skill` / `task` aren't forced into Chinese the way `project` / `autopilot` are**:
+
+- **`issue` / `task`**: dev teams talk in English. The Chinese candidates ("任务" — too vague, almost synonymous with "工作"; "工单" — IT ticket connotation; "议题" — GitHub-style but doesn't match the product feel) all read worse than `issue`. **But** in long-form doc prose, repeating lowercase `task` 50× breaks the rhythm — so prose is allowed to use `执行任务`, while UI strings and state names stay lowercase English.
+- **`skill`**: Multica-specific concept with no established Chinese term.
+- **`project` → "项目"**: settled mainstream Chinese word. Feishu / Tower / Teambition / PingCode / GitHub Projects — every Chinese product translates it. No product keeps `project` in Chinese context.
+- **`autopilot` → "自动化"**: in Chinese, "autopilot" associates with Tesla's "自动驾驶" and doesn't match what the feature does (run tasks on a schedule). Notion and Feishu both use "自动化"; that's the industry consensus.
+
+### Don't translate — brands and acronyms
+
+| Category | Terms |
+| --- | --- |
+| Brands | **Multica**, GitHub, Slack, Google, Anthropic, OpenAI, Claude, Codex, Cursor, Linear, Jira |
+| Acronyms | API, CLI, URL, SDK, OAuth, JWT, SSO, WebSocket, HTTP, JSON, YAML, SQL |
+
+### Translate fully — concepts
+
+| English | Chinese |
+| --- | --- |
+| Workspace | **工作区** |
+| Agent | **智能体** |
+| Project | **项目** |
+| Autopilot | **自动化** |
+| Daemon | **守护进程** |
+| Runtime | **运行时** |
+| Inbox | **收件箱** |
+| Comment | **评论** |
+| Reply | **回复** |
+| Notifications | **通知** |
+| Member | **成员** |
+| Label | **标签** |
+| Settings | **设置** |
+| Onboarding | **上手引导** |
+
+### Translate fully — generic UI words
+
+| English | Chinese |
+| --- | --- |
+| Invite / Invitation | 邀请 |
+| Search | 搜索 |
+| Email | 邮箱 (label) / 邮件 (action) |
+| Password | 密码 |
+| Sign in / Log in | 登录 |
+| Sign up | 注册 |
+| Sign out / Log out | 退出登录 |
+| Save / Cancel / Delete | 保存 / 取消 / 删除 |
+| Confirm / Continue / Back | 确认 / 继续 / 返回 |
+| Edit / New / Create / Add | 编辑 / 新建 / 创建 / 添加 |
+| Remove / Send / Open / Close | 移除 / 发送 / 打开 / 关闭 |
+| Done / Loading... | 完成 / 加载中... |
+| Profile / Account / Appearance | 个人资料 / 账号 / 外观 |
+| Theme / Language | 主题 / 语言 |
+| Light / Dark / System | 浅色 / 深色 / 跟随系统 |
+| Active / Archived | 活跃 (or 启用) / 已归档 |
+| Status / Priority | 状态 / 优先级 |
+| Assignee / Reporter | 负责人 / 报告人 |
+| Description / Title | 描述 / 标题 |
+| Date / Time | 日期 / 时间 |
+| Today / Yesterday / Tomorrow | 今天 / 昨天 / 明天 |
+| Empty / Failed / Success | 空 / 失败 / 成功 |
+| Error / Warning | 错误 / 警告 |
+
+### Roles and status enums (lowercase English, not translated)
+
+These are schema-level identifiers; render as lowercase English even in Chinese context.
+
+- Roles: `owner` / `admin` / `member`
+- Issue status: `backlog` / `todo` / `in_progress` / `in_review` / `done` / `blocked` / `cancelled`
+
+In UI, surface them in English (optionally `code-style` wrapped):
+
+- "你需要 owner 权限"
+- "已切换到 in_progress"
+
+### Word combination rules
+
+Always put **a single space** between an English word (entity / brand / acronym) and surrounding Chinese:
+
+- "Create new issue" → "新建 issue"
+- "Assign to agent" → "分配给智能体"
+- "Configure runtime" → "配置运行时"
+- "Stop daemon" → "停止守护进程"
+
+### Plurals and counts
+
+i18next uses `_one` / `_other`; Chinese has no grammatical number, only fill `_other`.
+
+```json
+// en/issues.json
+{
+  "issue_count_one": "{{count}} issue",
+  "issue_count_other": "{{count}} issues"
+}
+
+// zh-Hans/issues.json
+{
+  "issue_count_other": "{{count}} 个 issue"
+}
+```
+
+Common count formats:
+
+- `{{count}} issues` → `{{count}} 个 issue`
+- `{{count}} agents` → `{{count}} 个智能体`
+- `{{count}} workspaces` → `{{count}} 个工作区`
+- `{{count}} comments` → `{{count}} 条评论`
+- `{{count}} members` → `{{count}} 位成员`
+- `{{count}} skills` → `{{count}} 个 skill`
+
+### Interpolation
+
+Use `{{var}}`. Chinese translations may reorder for natural sentence flow.
+
+```json
+// en
+{ "welcome_message": "Welcome back, {{name}}!" }
+
+// zh-Hans
+{ "welcome_message": "欢迎回来，{{name}}！" }
+```
+
+### Translation key naming
+
+Three-level nesting: `feature.component.action`.
+
+```json
+{
+  "feature_or_component": {
+    "subcomponent_or_section": {
+      "action_or_label": "..."
+    }
+  }
+}
+```
+
+Examples:
+
+- `issues.toolbar.batch_update_success`
+- `issues.detail.comment_form.placeholder`
+- `inbox.empty.title`
+- `settings.preferences.language.title`
+
+### Web-only / desktop-only copy
+
+- Shared copy: top level of the namespace JSON
+- Web-only: `web` section
+- Desktop-only: `desktop` section
+
+See `auth.json` for the canonical example (the `web` section contains `prefer_desktop` / `desktop_handoff.*`).
+
+---
+
+## 3. Chinese voice and style
+
+### Punctuation
+
+- Full-width punctuation in Chinese: `，。：；！？`
+- Quotes: straight double quotes `"..."` to match the English source. Do not use `「」` or curly quotes.
+- Ellipsis: three dots `...` not the single character `…`. Match the English source.
+- Mixed Chinese-English: a single space on each side of the English word (see Word combination rules).
+
+### Style principles
+
+- **Concise and direct.** Avoid translation-ese: "对于 X 来说"、"作为 X"、"我们的"。
+- **Error messages**: gentle but clear. "无法保存修改" beats "保存修改失败了！".
+- **Buttons**: verb first, 2–4 characters. "取消"、"保存修改"、"立即同步".
+- **Tooltips**: full short sentence. "复制链接到剪贴板".
+- **Placeholders**: example-style. "输入 issue 标题...".
+
+### Where to look when in doubt
+
+When the glossary doesn't cover a term, look at:
+
+1. `apps/docs/content/docs/*.zh.mdx` — the de facto Chinese voice standard, 20+ pages of consistent translation
+2. `packages/views/locales/zh-Hans/auth.json` and `editor.json` — JSON structure + selector API patterns
+3. `packages/views/auth/login-page.tsx` — component-level selector API call site
+4. `packages/views/settings/components/preferences-tab.tsx` — language switcher reference
+
+---
+
+## Updating this page
+
+If you change a rule here, also:
+
+1. Apply it in the relevant locale JSONs / CLAUDE.md / docs page
+2. Note the change in the PR description so reviewers know to look for downstream sweep
+
+This page is the contract; nothing else overrides it.

apps/docs/content/docs/developers/conventions.zh.mdx

@@ -0,0 +1,301 @@
+---
+title: 规范
+description: 代码命名规范、i18n 翻译术语表、中文风格指南的唯一权威来源。
+---
+
+本页是代码命名规范、i18n 翻译术语表、中文风格指南的唯一权威来源。原本散落在 `packages/views/locales/glossary.md` 和各处注释里的规则现在都收拢到这里。
+
+写 Multica 代码、改翻译、写中文产品文案，都从这一页查。
+
+---
+
+## 1. 代码命名
+
+### 路由
+
+工作区前置路由（用户进入工作区之前能访问的路由）必须用单个单词，或者 `/{noun}/{verb}` 格式。
+
+- ✅ `/login`、`/inbox`、`/workspaces/new`
+- ❌ `/new-workspace`、`/create-team`、`/accept-invite`
+
+根目录的连字符词组会跟用户自选 workspace slug 冲突，逼着团队不停审保留字列表。把名词（`workspaces`）保留下来，整个 `/workspaces/*` 子树自动受保护。
+
+### 工作区路由
+
+永远用 `/{slug}/{section}` —— `/{slug}/issues`、`/{slug}/agents`、`/{slug}/settings`。共享代码不要复制路由逻辑，统一走 `useNavigation().push()`，不要直接用框架的 link API。
+
+### 包与模块
+
+monorepo 的包边界是硬约束：
+
+| 包 | 可依赖 | 不能依赖 |
+| --- | --- | --- |
+| `packages/core` | 仅平台无关基础库 | `react-dom`、`localStorage`、`process.env`、`next/*`、UI 库 |
+| `packages/ui` | 无业务依赖 | `@multica/core`、业务逻辑 |
+| `packages/views` | `core/`、`ui/` | `next/*`、`react-router-dom`、stores |
+| `apps/web/platform/` | `next/*` | 其他 app |
+| `apps/desktop/.../platform/` | `react-router-dom`、electron | 其他 app |
+
+两个 app 都有的逻辑，**必须**抽到共享包。"小段重复"也不算例外。
+
+### 文件与组件
+
+- 文件名：`kebab-case.tsx` / `kebab-case.ts`（如 `agent-row-actions.tsx`）
+- 组件：`PascalCase`（如 `AgentRowActions`）
+- Hook：`useCamelCase`（如 `useWorkspaceId`）
+- 测试：与源文件同目录，命名 `<file>.test.ts(x)`
+- Zustand store：`<feature>-store.ts`，导出名 `use<Feature>Store`
+
+### 数据库（Go + sqlc）
+
+- 表名：`snake_case` 单数（`user`、`workspace`、`agent_runtime`）
+- 字段：`snake_case`（`workspace_id`、`created_at`、`last_seen_at`）
+- 外键：`<table>_id`
+- 布尔：`is_<state>` 或者 `<state>_at`（状态变化优先用时间戳形式）
+- 迁移文件：`NNN_descriptive_name.up.sql` + `.down.sql`，**永远写双向**
+
+### Go
+
+- 标准 `gofmt` + `go vet`，无例外
+- Handler 文件按域命名：`agent.go`、`auth.go`、`runtime.go`
+- 测试：`<file>_test.go` 同目录
+- handler 里 UUID 解析遵守根 `CLAUDE.md` 的规则：边界输入用 `parseUUIDOrBadRequest`，可信回环用 `parseUUID`（panic 版），永远不要直接用 `util.ParseUUID` 不查 error
+
+### TypeScript
+
+- 网络上 API 响应是 `snake_case`，api client 在边界处转成 `camelCase`。**TS 代码内部一律 camelCase**
+- 类型：`PascalCase`（`Issue`、`AgentRuntime`），不加 `IPrefix`，不加 `_t` 后缀
+- 枚举：优先用 string literal union，需要 runtime 迭代时才用 `enum`
+- TanStack Query key：用 `<feature>/queries.ts` 里的工厂函数，例如 `issueKeys.detail(id)`
+
+### Issue 编号
+
+每个 issue 有人类可读的编号，比如 `MUL-123`：工作区 `issue_prefix`（3 个大写字母）+ 流水号。前缀在工作区创建时定，之后不可改。
+
+### 代码注释
+
+**只允许英文**。Go 和 TypeScript 都强制。如果在代码里看到中文注释，那就是 bug，替换掉。
+
+### Commit message
+
+Conventional 格式：`feat(scope)`、`fix(scope)`、`refactor(scope)`、`docs`、`test(scope)`、`chore(scope)`。按意图原子化分组。
+
+---
+
+## 2. i18n 翻译术语表
+
+这是每个翻译 PR 都必须遵守的术语表。原本在 `packages/views/locales/glossary.md`，那个文件现在是个 stub，指向这一页。
+
+### 核心区分：实体 vs 概念
+
+Multica 的产品名词分两类：
+
+- **实体（typed entity）** —— 有 URL、有数据库 row、是 API 响应里某种 type 的东西。中文里**用小写英文**呈现，视觉上像类型名，告诉读者"这是 Multica 系统里的特定实体"。
+- **概念（concept）** —— 不是数据库实体的普通名词。**完整翻译成中文**，CN 用户看不到生硬的英文。
+
+这套规则与 `apps/docs/content/docs/*.zh.mdx` 完全对齐 —— docs 是已经实战 20+ 篇的 CN voice 标准。
+
+### 实体词的混合规则（`issue` / `skill` / `task`）
+
+`issue` / `skill` / `task` 是 Multica 的核心实体。schema 字段、API 字段、产品 UI 标签都用英文。中文里采用**混合规则** —— 词出现在哪里决定怎么写：
+
+| 场景 | 写法 | 例 |
+| --- | --- | --- |
+| **UI 短句 / 状态名 / 代码上下文** | 小写英文 | "排队中的 task"、"创建子 issue"、"为智能体注入 skill" |
+| **doc 标题 / 章节标题** | 首字母大写英文，**或**对应中文术语 | "Issue 与 project"、"Skills"、"执行任务" |
+| **doc 正文长篇讨论中作为主语** | 中文术语，首次出现配括号英文 | "**执行任务**（task）是智能体每一次工作的单位" |
+| **API / DB 字段** | 永远 `task` / `issue` / `skill` | `task_id`、`issue_status`、`skill_uuid` |
+
+中文术语对照：
+
+- `task` ↔ `执行任务`（上下文清楚后可简写为「任务」）
+- `issue` 没有公认中文译法 —— 保留英文；标题可大写为 `Issue`
+- `skill` 没有公认中文译法 —— 保留英文；标题可大写为 `Skills`
+
+**为什么 `issue` / `skill` / `task` 不强制译，而 `project` / `autopilot` 必译**：
+
+- **`issue` / `task`**：dev 团队习惯说英文，"任务"在中文里和"工作"几乎同义太空泛，"工单"是 IT 工单语义，"议题"是 GitHub 风格但用户场景不匹配 —— 三个候选都不如 `issue` 准确。**但**在长篇 doc 正文里，重复 50 次 `task` 节奏不顺，所以正文允许用 `执行任务`，UI 短句、状态名仍保持小写英文。
+- **`skill`**：Multica 特有概念，没有公认中文译法。
+- **`project` 翻成「项目」**：中文里早就稳定的日常词。飞书 / Tower / Teambition / PingCode / GitHub Projects 中文版 0 例外都翻译成「项目」，没有产品保留 `project`。
+- **`autopilot` 翻成「自动化」**：autopilot 在中文里联想到特斯拉的「自动驾驶」，跟产品功能（按周期跑 task）对应不上。Notion / 飞书都用「自动化」，是行业共识。
+
+### 完整翻译 —— 概念词
+
+| 英 | 中 |
+| --- | --- |
+| Workspace | **工作区** |
+| Agent | **智能体** |
+| Project | **项目** |
+| Autopilot | **自动化** |
+| Daemon | **守护进程** |
+| Runtime | **运行时** |
+| Inbox | **收件箱** |
+| Comment | **评论** |
+| Reply | **回复** |
+| Notifications | **通知** |
+| Member | **成员** |
+| Label | **标签** |
+| Settings | **设置** |
+| Onboarding | **上手引导** |
+
+### 不翻 —— 品牌名 + 通用缩写
+
+| 类别 | 词 |
+| --- | --- |
+| 品牌 | **Multica**、GitHub、Slack、Google、Anthropic、OpenAI、Claude、Codex、Cursor、Linear、Jira |
+| 缩写 | API、CLI、URL、SDK、OAuth、JWT、SSO、WebSocket、HTTP、JSON、YAML、SQL |
+
+### 完整翻译 —— 通用 UI 词
+
+| 英 | 中 |
+| --- | --- |
+| Invite / Invitation | 邀请 |
+| Search | 搜索 |
+| Email | 邮箱（label）/ 邮件（action） |
+| Password | 密码 |
+| Sign in / Log in | 登录 |
+| Sign up | 注册 |
+| Sign out / Log out | 退出登录 |
+| Save / Cancel / Delete | 保存 / 取消 / 删除 |
+| Confirm / Continue / Back | 确认 / 继续 / 返回 |
+| Edit / New / Create / Add | 编辑 / 新建 / 创建 / 添加 |
+| Remove / Send / Open / Close | 移除 / 发送 / 打开 / 关闭 |
+| Done / Loading... | 完成 / 加载中... |
+| Profile / Account / Appearance | 个人资料 / 账号 / 外观 |
+| Theme / Language | 主题 / 语言 |
+| Light / Dark / System | 浅色 / 深色 / 跟随系统 |
+| Active / Archived | 活跃（或 启用）/ 已归档 |
+| Status / Priority | 状态 / 优先级 |
+| Assignee / Reporter | 负责人 / 报告人 |
+| Description / Title | 描述 / 标题 |
+| Date / Time | 日期 / 时间 |
+| Today / Yesterday / Tomorrow | 今天 / 昨天 / 明天 |
+| Empty / Failed / Success | 空 / 失败 / 成功 |
+| Error / Warning | 错误 / 警告 |
+
+### 角色名 + 状态名（小写英文，不翻）
+
+这些是 schema-level 标识符，中文环境也保持小写英文：
+
+- 角色：`owner` / `admin` / `member`
+- Issue 状态：`backlog` / `todo` / `in_progress` / `in_review` / `done` / `blocked` / `cancelled`
+
+UI 里展示这些值时保持英文（必要时用 code-style 包起来）：
+
+- "你需要 owner 权限"
+- "已切换到 in_progress"
+
+### 词组组合规则
+
+英文词（实体名 + 品牌名 + 缩写）与中文之间**加单空格**：
+
+- "Create new issue" → "新建 issue"
+- "Assign to agent" → "分配给智能体"
+- "Configure runtime" → "配置运行时"
+- "Stop daemon" → "停止守护进程"
+
+### 复数与计数
+
+i18next 用 `_one` / `_other`；中文不区分语法单复数，只填 `_other`。
+
+```json
+// en/issues.json
+{
+  "issue_count_one": "{{count}} issue",
+  "issue_count_other": "{{count}} issues"
+}
+
+// zh-Hans/issues.json
+{
+  "issue_count_other": "{{count}} 个 issue"
+}
+```
+
+常见计数格式：
+
+- `{{count}} issues` → `{{count}} 个 issue`
+- `{{count}} agents` → `{{count}} 个智能体`
+- `{{count}} workspaces` → `{{count}} 个工作区`
+- `{{count}} comments` → `{{count}} 条评论`
+- `{{count}} members` → `{{count}} 位成员`
+- `{{count}} skills` → `{{count}} 个 skill`
+
+### 插值
+
+用 `{{var}}` 形式。中文翻译可以调整位置以符合中文语序。
+
+```json
+// en
+{ "welcome_message": "Welcome back, {{name}}!" }
+
+// zh-Hans
+{ "welcome_message": "欢迎回来，{{name}}！" }
+```
+
+### Key 命名约定
+
+3 层嵌套：`feature.component.action`。
+
+```json
+{
+  "feature_or_component": {
+    "subcomponent_or_section": {
+      "action_or_label": "..."
+    }
+  }
+}
+```
+
+实例：
+
+- `issues.toolbar.batch_update_success`
+- `issues.detail.comment_form.placeholder`
+- `inbox.empty.title`
+- `settings.preferences.language.title`
+
+### Web-only / Desktop-only 文案位置
+
+- 共享文案：放 namespace JSON 顶层
+- Web-only：放 `web` 段
+- Desktop-only：放 `desktop` 段
+
+参考 `auth.json`（`web` 段含 `prefer_desktop` / `desktop_handoff.*`）。
+
+---
+
+## 3. 中文风格
+
+### 标点
+
+- 中文用全角标点：`，。：；！？`
+- 引号：用 `"..."`（直引号），与英文 source 保持一致。**不要**用 `「」` 或弯引号
+- 省略号：用 `...`（三点）而非 `…`（单字符），与英文 source 保持一致
+- 中英混排：英文词左右各加 1 个空格（详见词组组合规则）
+
+### 风格原则
+
+- **简洁直白**：避免翻译腔，"对于 X 来说"、"作为 X"、"我们的"
+- **错误信息**：温和但明确，"无法保存修改" 优于 "保存修改失败了！"
+- **按钮**：动词开头，2-4 字最佳。"取消"、"保存修改"、"立即同步"
+- **Tooltip**：完整短句。"复制链接到剪贴板"
+- **placeholder**：示例性提示。"输入 issue 标题..."
+
+### 拿不准的时候去哪查
+
+术语表没覆盖的词，按这个顺序查：
+
+1. `apps/docs/content/docs/*.zh.mdx` —— CN voice 事实标准，20+ 篇高度一致
+2. `packages/views/locales/zh-Hans/auth.json` 和 `editor.json` —— JSON 结构 + selector API 用法参考
+3. `packages/views/auth/login-page.tsx` —— 组件层 selector API 调用参考
+4. `packages/views/settings/components/preferences-tab.tsx` —— 语言切换器参考
+
+---
+
+## 修改这一页时
+
+改本页规则的同时还要：
+
+1. 把规则在相关 locale JSON / CLAUDE.md / docs 页面里同步落地
+2. PR 描述里写明改了什么，方便 reviewer 检查下游是否跟着改了
+
+本页是契约，其他文档不能 override。

apps/docs/content/docs/developers/meta.json

@@ -0,0 +1,4 @@
+{
+  "title": "Developers",
+  "pages": ["conventions"]
+}

apps/docs/content/docs/developers/meta.zh.json

@@ -1,4 +1,4 @@
 {
   "title": "Developers",
-  "pages": ["contributing", "architecture"]
+  "pages": ["contributing", "architecture", "conventions"]
 }

apps/docs/content/docs/environment-variables.mdx

@@ -66,13 +66,19 @@ Multica stores user-uploaded attachments (images and files in comments). **S3 is
 
 | Variable | Default | Description |
 |---|---|---|
-| `S3_BUCKET` | empty | Setting this enables S3 storage |
-| `S3_REGION` | `us-west-2` | AWS region |
+| `S3_BUCKET` | empty | **Bucket name only** (for example `my-bucket`). Do **not** include the `.s3.<region>.amazonaws.com` suffix — the server constructs the public host from `S3_BUCKET` + `S3_REGION`. Setting this enables S3 storage |
+| `S3_REGION` | `us-west-2` | AWS region. Must match the bucket's actual region — it is used both for SDK signing and for building the public URL |
 | `AWS_ACCESS_KEY_ID` / `AWS_SECRET_ACCESS_KEY` | empty | Static credentials. When both are unset, the AWS SDK default credential chain is used (IAM role / environment credentials) |
 | `AWS_ENDPOINT_URL` | empty | Custom S3-compatible endpoint (for example [MinIO](https://min.io/)). Setting this switches to path-style URLs |
 
 **When `S3_BUCKET` is unset**: the server logs `"S3_BUCKET not set, cloud upload disabled"` at startup, and all uploads fall back to local disk.
 
+**Public URLs** are constructed in this order of priority:
+
+1. `https://<CLOUDFRONT_DOMAIN>/<key>` if `CLOUDFRONT_DOMAIN` is set.
+2. `<AWS_ENDPOINT_URL>/<S3_BUCKET>/<key>` (path-style) if `AWS_ENDPOINT_URL` is set.
+3. `https://<S3_BUCKET>.s3.<S3_REGION>.amazonaws.com/<key>` (virtual-hosted-style). When `S3_BUCKET` contains dots, the server falls back to `https://s3.<S3_REGION>.amazonaws.com/<S3_BUCKET>/<key>` (path-style) because the AWS-issued wildcard TLS certificate does not validate dotted bucket hosts.
+
 ### Local disk (when S3 is not configured)
 
 | Variable | Default | Description |

apps/docs/content/docs/environment-variables.zh.mdx

@@ -66,13 +66,19 @@ Multica 存储用户上传的附件（评论里的图片、文件等）。**优
 
 | 环境变量 | 默认值 | 说明 |
 |---|---|---|
-| `S3_BUCKET` | 空 | 设了就启用 S3 存储 |
-| `S3_REGION` | `us-west-2` | AWS 区域 |
+| `S3_BUCKET` | 空 | **只填 bucket 名**（例如 `my-bucket`），**不要**带 `.s3.<region>.amazonaws.com` 后缀——server 会用 `S3_BUCKET` + `S3_REGION` 自己拼公开 host。设了就启用 S3 存储 |
+| `S3_REGION` | `us-west-2` | AWS 区域。必须和 bucket 所在区域一致——SDK 签名和公开 URL 都用它 |
 | `AWS_ACCESS_KEY_ID` / `AWS_SECRET_ACCESS_KEY` | 空 | 静态凭证。全未设时用 AWS SDK 默认凭证链（IAM role / 环境凭证）|
 | `AWS_ENDPOINT_URL` | 空 | 自定义 S3 兼容端点（例如 [MinIO](https://min.io/)）。设了会切到 path-style URL |
 
 **`S3_BUCKET` 未设时**：server 启动时打 info 日志 `"S3_BUCKET not set, cloud upload disabled"`，所有上传回落到本地磁盘。
 
+**公开 URL** 按优先级拼装：
+
+1. 设了 `CLOUDFRONT_DOMAIN` → `https://<CLOUDFRONT_DOMAIN>/<key>`
+2. 设了 `AWS_ENDPOINT_URL` → `<AWS_ENDPOINT_URL>/<S3_BUCKET>/<key>`（path-style）
+3. 默认走 AWS S3 → `https://<S3_BUCKET>.s3.<S3_REGION>.amazonaws.com/<key>`（virtual-hosted-style）。bucket 名含点时会回落到 `https://s3.<S3_REGION>.amazonaws.com/<S3_BUCKET>/<key>`（path-style），因为 AWS 通配证书无法覆盖含点 host。
+
 ### 本地磁盘（S3 未配时）
 
 | 环境变量 | 默认值 | 说明 |

apps/docs/content/docs/getting-started/self-hosting.zh.mdx

@@ -212,13 +212,15 @@ Changes take effect after restarting the backend / compose stack. The web UI rea
 
 ### File Storage (Optional)
 
-For file uploads and attachments, configure S3 and CloudFront:
+For file uploads and attachments, configure S3 and (optionally) CloudFront:
 
 | Variable | Description |
 |----------|-------------|
-| `S3_BUCKET` | S3 bucket name |
-| `S3_REGION` | AWS region (default: `us-west-2`) |
-| `CLOUDFRONT_DOMAIN` | CloudFront distribution domain |
+| `S3_BUCKET` | Bucket name only (e.g. `my-bucket`). Do **not** include the `.s3.<region>.amazonaws.com` suffix — the server constructs the public URL from `S3_BUCKET` + `S3_REGION` |
+| `S3_REGION` | AWS region (default: `us-west-2`). Must match the bucket's actual region — used for both SDK signing and public URLs |
+| `AWS_ACCESS_KEY_ID` / `AWS_SECRET_ACCESS_KEY` | Static credentials. When both are unset, the AWS SDK default credential chain is used |
+| `AWS_ENDPOINT_URL` | Custom S3-compatible endpoint (e.g. MinIO, R2, B2). Setting this switches the public URL to path-style |
+| `CLOUDFRONT_DOMAIN` | CloudFront distribution domain — when set, public URLs use this host instead of the S3 host |
 | `CLOUDFRONT_KEY_PAIR_ID` | CloudFront key pair ID for signed URLs |
 | `CLOUDFRONT_PRIVATE_KEY` | CloudFront private key (PEM format) |
 

apps/docs/content/docs/guides/agents.zh.mdx

@@ -40,7 +40,7 @@ The daemon auto-detects which CLIs are available on your PATH and registers them
 
 Multica supports two layers of skills:
 
-- **Local skills** — Skills already installed in your local runtime (e.g., `.claude/skills/`, `.config/opencode/skills/`) are automatically discovered and used by agents. You do **not** need to upload them to Multica.
+- **Local skills** — Skills already installed in your local runtime (e.g., `.claude/skills/`, `.opencode/skills/`) are automatically discovered and used by agents. You do **not** need to upload them to Multica.
 - **Workspace skills** — Skills created or imported in the Multica Skills page are shared across the workspace. They are automatically injected into agent runs as supplementary context, so every team member's agents benefit from them.
 
 Workspace skills are designed for team-wide sharing and collaboration — codify your team's best practices once, and every agent can leverage them:

apps/docs/content/docs/meta.json

@@ -9,7 +9,9 @@
     "workspaces",
     "members-roles",
     "issues",
+    "projects",
     "comments",
+    "project-resources",
     "---Agents---",
     "agents",
     "agents-create",
@@ -32,6 +34,8 @@
     "---Reference---",
     "cli",
     "auth-tokens",
-    "desktop-app"
+    "desktop-app",
+    "---Developers---",
+    "developers"
   ]
 }

apps/docs/content/docs/meta.zh.json

@@ -9,6 +9,7 @@
     "workspaces",
     "members-roles",
     "issues",
+    "projects",
     "comments",
     "---智能体---",
     "agents",
@@ -32,6 +33,8 @@
     "---参考---",
     "cli",
     "auth-tokens",
-    "desktop-app"
+    "desktop-app",
+    "---开发者---",
+    "developers"
   ]
 }

apps/docs/content/docs/project-resources.mdx

@@ -0,0 +1,144 @@
+---
+title: Project Resources
+description: Attach typed pointers (Git repos today, more later) to a project so agents can pick them up as scoped context.
+---
+
+A **Project Resource** is a typed pointer — a Git repo URL today, a Notion page or document link tomorrow — attached to a [project](/workspaces). When an [agent](/agents) runs against an issue inside that project, the daemon automatically writes the project's resource list into the agent's working directory and into its [meta-skill](/skills) prompt.
+
+The result: the agent knows which repo to check out, which docs are the "primary references" for this project, without anyone copy-pasting context into the issue body.
+
+## Mental model
+
+A project is no longer just a label. It is a small **resource container**:
+
+- A project has 0..N **resources**.
+- A resource has a `resource_type` (e.g. `github_repo`) and a `resource_ref` (a JSON payload typed by `resource_type`).
+- New resource types add a string + a handler. **No schema migration. No frontend rewrite.**
+
+This shape is intentional — it's the same pattern Multica already uses for agent providers: a `type` discriminator and a typed payload. It keeps the schema stable so adding "Notion page", "Google Doc", "uploaded file", or "external URL" later is a small, additive change.
+
+## Today: `github_repo`
+
+The first resource type ships ready to use:
+
+```json
+{
+  "resource_type": "github_repo",
+  "resource_ref": {
+    "url": "https://github.com/owner/repo",
+    "default_branch_hint": "main"
+  }
+}
+```
+
+`default_branch_hint` is optional — if present, the daemon surfaces it in the meta-skill so the agent knows which branch to base its work on.
+
+## Attaching repos at project creation
+
+In the **Web** or **Desktop** app, opening *New project* now shows a **Repos** pill alongside Status / Priority / Lead. Selecting workspace-bound repos (or pasting an ad-hoc URL) attaches them as `github_repo` resources the moment the project is created.
+
+From the **CLI**:
+
+```bash
+# Create + attach in one shot. The server attaches resources in the same
+# transaction as the project create — invalid resources roll back the whole
+# operation, so you never end up with a project that has half its resources.
+multica project create \
+  --title "Agent UX 2026" \
+  --repo https://github.com/multica-ai/multica
+
+# Manage resources later
+multica project resource list <project-id>
+multica project resource add  <project-id> --type github_repo --url <url>
+multica project resource remove <project-id> <resource-id>
+
+# Generic escape hatch for any resource_type the server understands —
+# no CLI change needed when a new type ships:
+multica project resource add <project-id> \
+  --type notion_page \
+  --ref '{"page_id":"…","title":"…"}'
+```
+
+`--repo` may be repeated; each value is attached as a separate `github_repo` resource.
+
+## What the agent sees at runtime
+
+When the daemon spawns an agent for an issue inside a project, two things happen:
+
+### 1. `.multica/project/resources.json`
+
+A structured pass-through of the API response, written into the agent's working directory:
+
+```json
+{
+  "project_id": "…",
+  "project_title": "Agent UX 2026",
+  "resources": [
+    {
+      "id": "…",
+      "resource_type": "github_repo",
+      "resource_ref": {
+        "url": "https://github.com/multica-ai/multica",
+        "default_branch_hint": "main"
+      }
+    }
+  ]
+}
+```
+
+Skills, helper scripts, or the agent itself can parse this file when they need the *exact* set of resources for the run.
+
+### 2. A "Project Context" section in the meta-skill prompt
+
+The agent's `CLAUDE.md` / `AGENTS.md` (depending on provider) now includes a human-readable summary:
+
+```
+## Project Context
+
+This issue belongs to **Agent UX 2026**.
+
+Project resources (also written to `.multica/project/resources.json`):
+
+- **GitHub repo**: https://github.com/multica-ai/multica (default branch: `main`)
+
+Resources are pointers — open them only when relevant to the task. For
+`github_repo` resources, use `multica repo checkout <url>` to fetch the code.
+```
+
+The text is intentionally minimal. The full payload is on disk; the prompt only orients the agent so it knows the project exists and what's attached.
+
+### Failure mode
+
+Resource fetch is **best-effort**. If the API call fails, the project section is omitted from the prompt and the file is not written, but the task still starts. Agents never block on missing project context.
+
+## Adding a new resource type
+
+The whole point of the abstraction is that new types are cheap. The full path:
+
+1. **Server validator** (`server/internal/handler/project_resource.go`) — add a case in `validateAndNormalizeResourceRef` that parses and normalizes the new payload.
+2. **Daemon meta-skill formatter** (`server/internal/daemon/execenv/runtime_config.go`) — add a case in `formatProjectResource` so the agent prompt renders the new type as a readable bullet.
+3. **TypeScript types** (`packages/core/types/project.ts`) — extend `ProjectResourceType` and add the payload interface.
+4. **UI renderer** (`packages/views/projects/components/project-resources-section.tsx`) — add a case in `ResourceRow` for the new type.
+
+There is **no schema migration**, no new sqlc query, no new endpoint, **and no CLI change** — the CLI's generic `--ref '<json>'` flag accepts any payload the validator understands, so day-one support for a new type is purely the four steps above. (You may *optionally* add a per-type CLI shortcut later; not required.)
+
+The same `project_resource` table and the same three CRUD calls handle every type.
+
+## Workspace repos vs. project repos
+
+The repo list shown to the agent (`## Repositories` block in `CLAUDE.md` / `AGENTS.md`) is chosen by the daemon claim handler with this precedence:
+
+- **Project has at least one `github_repo` resource** → only those repos are surfaced to the agent. Workspace-bound repos are intentionally hidden so the agent doesn't have to guess which one belongs to this issue.
+- **Project has no `github_repo` resources (or the issue isn't in a project)** → fall back to the workspace's repo list as before.
+
+This keeps the agent's working set tight: when a project is explicit about its repos, that's the authoritative answer. The structured resource list at `.multica/project/resources.json` always carries the full set, so a skill that wants to inspect everything still can.
+
+The daemon mirrors this on the checkout side: when a task arrives with project-scoped `github_repo` URLs, those URLs are merged into the per-workspace allowlist *and* synced into the local repo cache before the agent spawns. So a project repo URL that isn't bound at the workspace level is still a valid argument to `multica repo checkout` — the daemon won't reject it as "not configured." The allowlist split is internal: workspace-bound URLs and task-scoped URLs are tracked separately, so a workspace-repos refresh doesn't accidentally revoke a project URL mid-run.
+
+## What's intentionally **not** in scope here
+
+- **Cross-project sharing.** Each resource lives on exactly one project today.
+- **Per-skill resource scoping.** All resources are visible to every skill on the agent's run; type-aware filtering is a follow-up.
+- **Caching / sync.** `github_repo` is just metadata — checkout still happens via `multica repo checkout` on demand. Cached document text for Notion / Google Docs will arrive with those types.
+
+These are deliberate omissions — the goal of the first cut is to validate the abstraction with the smallest set of moving parts.

apps/docs/content/docs/projects.mdx

@@ -0,0 +1,49 @@
+---
+title: Projects
+description: Group related issues and track them as one unit — with priority, status, progress, and an owner.
+---
+
+import { Callout } from "fumadocs-ui/components/callout";
+
+A **project** in Multica is a container for related [issues](/issues). Use it when a body of work is bigger than one issue but smaller than a full workspace — a launch, a migration, a feature with multiple parts, an investigation that branches into several threads.
+
+Each project has a name, an icon, a description, a **lead** (a member or an [agent](/agents)), a **status** (`planned` / `in_progress` / `paused` / `completed` / `cancelled`), a **priority** (`urgent` / `high` / `medium` / `low` / `none`), and a **progress** percentage that's auto-derived from the status of its linked issues.
+
+## How projects relate to issues
+
+Projects and issues are independent objects with a many-to-one relationship: an issue can belong to **at most one** project; a project holds **any number of** issues. Linking and unlinking is reversible at any time — drag in the board view, or use the project picker on the issue's right-side properties panel.
+
+The progress bar on a project is computed from its linked issues — the more issues hit `done`, the further it fills. Issues that are `cancelled` are excluded from the count; issues in `backlog` count toward the denominator but not the numerator.
+
+## Pinning to the sidebar
+
+Click the pin icon in a project's top-right corner to add it to your sidebar's pinned list. Pinned projects stay one click away no matter where you are in the workspace; everyone on the team can pin independently — pins are personal.
+
+The sidebar **Workspace → Projects** link always shows every project in the workspace; pinning is a personal shortcut on top of that.
+
+## Attaching resources
+
+Each project has a **Resources** section where you attach GitHub repositories. Once attached, any [agent](/agents) assigned to issues in this project can read and write to those repos when executing tasks — Multica passes the repo URLs as context to the [daemon](/daemon-runtimes).
+
+Resources are per-project; if multiple projects share a repo, attach it to each one.
+
+## Deleting a project
+
+Deleting a project **does not delete its issues**. The linked issues are simply unlinked and revert to the workspace's flat issue list. This is intentional — work that was scoped to a project is rarely throwaway, even when the framing of the project changes.
+
+<Callout type="info">
+If you want to delete the work too, archive or delete the issues first, then delete the project.
+</Callout>
+
+## Project lead
+
+The lead is the person — or agent — accountable for the project. It's a soft signal, not an access control: any workspace member can edit a project regardless of who's lead. A project's lead can be:
+
+- A workspace member (human teammate)
+- An [agent](/agents) — useful when the project's work is mostly delegated to an agent (e.g., "Weekly bug triage" led by a triage agent)
+
+## Next
+
+- [Issues](/issues) — the unit of work that lives inside projects
+- [Agents as project lead](/agents) — when an agent is the right owner
+- [How Multica works](/how-multica-works) — the broader picture

apps/docs/content/docs/projects.zh.mdx

@@ -0,0 +1,49 @@
+---
+title: 项目
+description: 把相关的 issue 归为一组当成一个单元来跟进 —— 有优先级、状态、进度和负责人。
+---
+
+import { Callout } from "fumadocs-ui/components/callout";
+
+Multica 里的**项目**（project）是相关 [issue](/issues) 的容器。当一摊工作比单个 issue 大、又比整个工作区小的时候用它 —— 一次发布、一次迁移、一个分多块做的功能、一个会拆出多个线索的调研。
+
+每个项目有名字、图标、描述、**负责人**（lead，可以是成员，也可以是 [智能体](/agents)）、**状态**（`planned` / `in_progress` / `paused` / `completed` / `cancelled`）、**优先级**（`urgent` / `high` / `medium` / `low` / `none`），以及一个根据关联 issue 状态自动算出来的**进度**百分比。
+
+## 项目和 issue 的关系
+
+项目和 issue 是独立对象，多对一关系：一个 issue **最多属于一个**项目；一个项目可以容纳**任意多个** issue。关联和解除关联随时可逆 —— 在看板视图里拖动，或者在 issue 右侧 properties 面板用项目选择器。
+
+项目的进度条是按关联 issue 状态自动算出来的 —— 越多 issue 到 `done`，进度条越满。`cancelled` 的 issue 不计入分母；`backlog` 的 issue 计入分母但不计入分子。
+
+## pin 到侧边栏
+
+点项目右上角的 pin 图标，可以把这个项目加到侧边栏的固定区。pin 过的项目无论你在工作区哪里都一键可达；每个人独立 pin —— pin 是个人偏好。
+
+侧边栏 **Workspace → Projects** 链接始终展示工作区里所有项目；pin 只是在这之上的个人快捷方式。
+
+## 关联 resources
+
+每个项目有一个 **Resources** 区，可以挂 GitHub 仓库。挂上之后，被分配到这个项目里 issue 的 [智能体](/agents) 在执行 task 时可以读写这些仓库 —— Multica 会把仓库 URL 作为上下文传给 [守护进程](/daemon-runtimes)。
+
+Resources 是项目级别的；多个项目要共享同一个仓库，要分别挂上。
+
+## 删除项目
+
+删除项目**不会**删除它的 issue。关联的 issue 只是解除关联，回到工作区的扁平 issue 列表。这是刻意的 —— 即使项目本身的框架变了，里面的工作通常也不会是一次性的。
+
+<Callout type="info">
+如果你确实想把工作也删掉，先归档或删除 issue，再删除项目。
+</Callout>
+
+## 项目负责人
+
+负责人是为这个项目负总责的人 —— 或者智能体。这是一个软信号，不是权限控制：工作区任何成员都可以编辑项目，不管谁是负责人。项目负责人可以是：
+
+- 工作区里的成员（人）
+- [智能体](/agents) —— 当项目里的工作大部分要交给智能体时合适（例如"每周 bug 巡检"由一个巡检智能体担任 lead）
+
+## 下一步
+
+- [Issues](/issues) —— 项目里装的工作单元
+- [智能体担任项目负责人](/agents) —— 什么时候由智能体当 lead 合适
+- [Multica 怎么运转](/how-multica-works) —— 整体视图

apps/docs/content/docs/providers.mdx

@@ -21,7 +21,7 @@ For guidance on picking a tool when creating an agent, see [Creating and configu
 | **Hermes** | Nous Research | ✅ | ❌ | `.agent_context/skills/` (fallback) | Dynamic discovery |
 | **Kimi** | Moonshot | ✅ | ❌ | `.kimi/skills/` | Dynamic discovery |
 | **Kiro CLI** | Amazon | ✅ | ❌ | `.kiro/skills/` | Dynamic discovery |
-| **OpenCode** | SST | ✅ | ❌ | `.config/opencode/skills/` | Dynamic discovery |
+| **OpenCode** | SST | ✅ | ❌ | `.opencode/skills/` | Dynamic discovery |
 | **OpenClaw** | Open source | ✅ | ❌ | `.agent_context/skills/` (fallback) | Bound to the agent, can't be switched per task |
 | **Pi** | Inflection AI | ✅ (session is a file path) | ❌ | `.pi/skills/` | Dynamic discovery |
 
@@ -103,7 +103,7 @@ Each tool uses **its own** skill discovery path. Before a task runs, the Multica
 | Cursor | `.cursor/skills/` | ✅ Native |
 | Kimi | `.kimi/skills/` | ✅ Native |
 | Kiro CLI | `.kiro/skills/` | ✅ Native |
-| OpenCode | `.config/opencode/skills/` | ✅ Native |
+| OpenCode | `.opencode/skills/` | ✅ Native |
 | Pi | `.pi/skills/` | ✅ Native |
 | Gemini | `.agent_context/skills/` | ⚠️ Generic fallback |
 | Hermes | `.agent_context/skills/` | ⚠️ Generic fallback |

apps/docs/content/docs/providers.zh.mdx

@@ -21,7 +21,7 @@ Multica 内置支持 **11 款 AI 编程工具**。它们都实现了同一套接
 | **Hermes** | Nous Research | ✅ | ❌ | `.agent_context/skills/` （fallback）| 动态发现 |
 | **Kimi** | Moonshot | ✅ | ❌ | `.kimi/skills/` | 动态发现 |
 | **Kiro CLI** | Amazon | ✅ | ❌ | `.kiro/skills/` | 动态发现 |
-| **OpenCode** | SST | ✅ | ❌ | `.config/opencode/skills/` | 动态发现 |
+| **OpenCode** | SST | ✅ | ❌ | `.opencode/skills/` | 动态发现 |
 | **OpenClaw** | 开源项目 | ✅ | ❌ | `.agent_context/skills/` （fallback）| 绑定在智能体上，不能在任务里切换 |
 | **Pi** | Inflection AI | ✅（session 为文件路径）| ❌ | `.pi/skills/` | 动态发现 |
 
@@ -103,7 +103,7 @@ Inflection AI 出品，极简主义。**会话恢复机制特殊**——session
 | Cursor | `.cursor/skills/` | ✅ 原生 |
 | Kimi | `.kimi/skills/` | ✅ 原生 |
 | Kiro CLI | `.kiro/skills/` | ✅ 原生 |
-| OpenCode | `.config/opencode/skills/` | ✅ 原生 |
+| OpenCode | `.opencode/skills/` | ✅ 原生 |
 | Pi | `.pi/skills/` | ✅ 原生 |
 | Gemini | `.agent_context/skills/` | ⚠️ 通用 fallback |
 | Hermes | `.agent_context/skills/` | ⚠️ 通用 fallback |

apps/docs/content/docs/self-host-quickstart.mdx

@@ -116,4 +116,4 @@ Same flow as Cloud — see [Cloud quickstart → Steps 5-6](/cloud-quickstart#5-
 - [Environment variables](/environment-variables) — full env reference
 - [Auth setup](/auth-setup) — Resend / OAuth / signup allowlist in detail
 - [Troubleshooting](/troubleshooting) — start here when things go wrong
-- [Desktop app](/desktop-app) — released Desktop builds connect to Multica Cloud only; using Desktop with self-host requires a custom build (see the callout in the desktop-app page)
+- [Desktop app](/desktop-app) — optional Desktop setup via `~/.multica/desktop.json`; the web frontend + CLI remains the quickest self-host path

apps/docs/content/docs/self-host-quickstart.zh.mdx

@@ -115,4 +115,4 @@ multica setup self-host
 - [环境变量](/environment-variables) —— 完整 env 清单
 - [登录与注册配置](/auth-setup) —— Resend / OAuth / 注册白名单详细配置
 - [故障排查](/troubleshooting) —— 遇到问题先来这里
-- [桌面应用](/desktop-app) —— 发布版 Desktop 只连 Multica Cloud；要让 Desktop 连自部署后端需要自行构建（详见 desktop-app 页的提示）
+- [桌面应用](/desktop-app) —— 可以通过 `~/.multica/desktop.json` 连接 Desktop；Web 前端 + CLI 仍然是最快的自部署路径

apps/docs/content/docs/tasks.mdx

@@ -63,11 +63,13 @@ Automatic retry also has two extra conditions:
 
 <Callout type="warning">
 **Autopilot tasks don't retry automatically** by design. An Autopilot has its own firing cadence (e.g. daily); automatic retries on failure would overlap with the next scheduled run. If you need an immediate re-run after failure, use a manual rerun (next section).
+
+**How you'll know an Autopilot task failed**: a notification lands in your [Inbox](/inbox), and the associated issue's status reverts from `in_progress` back to `todo`. The [Autopilots](/autopilots) page also shows the latest run result per autopilot.
 </Callout>
 
 ## Manual rerun vs. automatic retry
 
-A **manual rerun** is one you trigger from the UI or CLI:
+A **manual rerun** is one you trigger from the CLI or the API (`POST /api/issues/{id}/rerun`):
 
 ```bash
 multica issue rerun <issue-id>
@@ -75,9 +77,10 @@ multica issue rerun <issue-id>
 
 Behavior:
 
-- **Cancels** the currently running task (if any)
-- Creates a **brand-new** task — attempt count resets to 1, even if the original task hit the attempt ceiling
-- Inherits the previous session ID; if the corresponding AI coding tool supports session resumption, the new task continues from the previous context
+- Targets the issue's **current agent assignee** — not whoever ran the most recent task. If the assignee changed since the last run, rerun follows the current assignment. To rerun a specific agent that is no longer the assignee, reassign the issue first, then rerun.
+- **Cancels** the assignee's queued or running task on this issue (if any). Tasks owned by other agents on the same issue (e.g. parallel @-mention runs) are left alone.
+- Creates a **brand-new** task — attempt count resets to 1, even if the original task hit the attempt ceiling.
+- Starts a **fresh agent session** — the prior session ID is **not** inherited. A manual rerun means you've judged the previous output bad, so resuming the same conversation would replay the same poisoned state. (Automatic retry, by contrast, does inherit the session — that path is for infrastructure failures, not bad output.)
 
 Comparison:
 
@@ -85,8 +88,9 @@ Comparison:
 |---|---|---|
 | Trigger | System, based on failure reason | You, manually |
 | Ceiling | 2 attempts | No limit |
-| Applicable sources | Issues, chat | All sources |
-| Session inheritance | Yes | Yes |
+| Applicable sources | Issues, chat | Issues with an agent assignee |
+| Agent picked | Same agent as the failed task | Issue's current assignee |
+| Session inheritance | Yes (resumes prior session) | No (fresh session) |
 
 ## How a failed task affects issue status
 
@@ -96,7 +100,7 @@ If an issue-triggered task fails (and no automatic retry succeeds) because the i
 
 Yes — as long as the AI coding tool supports session resumption.
 
-Multica pins the session ID **twice** during a task: once at the start (when the AI tool returns its first system message), and once at the end (on completion or failure). The first lets the daemon recover if it crashes mid-run; the second is reserved for future reruns. On the next rerun or automatic retry, that ID is passed back so the agent can pick up the previous conversation and file state.
+Multica pins the session ID **twice** during a task: once at the start (when the AI tool returns its first system message), and once at the end (on completion or failure). The first lets the daemon recover if it crashes mid-run; the second is reserved for the next **automatic retry**, where that ID is passed back so the agent can pick up the previous conversation and file state. **Manual rerun deliberately skips this** and starts a fresh session — see [Manual rerun vs. automatic retry](#manual-rerun-vs-automatic-retry).
 
 But **which AI coding tools actually support this** varies a lot:
 

apps/docs/content/docs/tasks.zh.mdx

@@ -63,11 +63,13 @@ Multica 服务器每 30 秒扫描一次，有两种超时会触发失败：
 
 <Callout type="warning">
 **Autopilots 任务不自动重试**是刻意设计。Autopilot 有自己的触发周期（例如每天一次）；如果失败又自动重试，会和下一个周期的任务重叠。需要失败后立即重跑，用手动重跑（下一节）。
+
+**怎么知道 Autopilot 失败了**：失败的 Autopilot 任务会在你的 [收件箱](/inbox) 里出现一条通知，关联的 issue 状态也会从 `in_progress` 退回 `todo`。直接打开 [Autopilots](/autopilots) 页面也能看到每条 autopilot 的最近运行结果。
 </Callout>
 
 ## 手动重跑和自动重试的区别
 
-**手动重跑**（rerun）是你从 UI 或命令行主动发起的：
+**手动重跑**（rerun）是你通过命令行或 API（`POST /api/issues/{id}/rerun`）主动发起的：
 
 ```bash
 multica issue rerun <issue-id>
@@ -75,9 +77,10 @@ multica issue rerun <issue-id>
 
 行为：
 
-- **取消**当前正在跑的任务（如果有）
-- 创建一个**全新**的执行任务——尝试次数重置为 1，即使原任务已达最大尝试
-- 继承上一次的会话 ID；如果对应的 AI 编程工具支持会话恢复，会接着上次的上下文继续
+- 跑的是 issue **当前的智能体分配人**——不是上一次跑过的 agent。如果分配人在上次运行后改了，rerun 会跟着新的分配人走。要重跑一个已经不再是分配人的智能体，先把 issue 改派回它，再 rerun。
+- **取消**该分配人在这条 issue 上 queued / running 的任务（如果有）。同 issue 上其它 agent 的任务（例如 @-mention 触发的并行任务）不会被一起取消。
+- 创建一个**全新**的执行任务——尝试次数重置为 1，即使原任务已达最大尝试。
+- 启动**全新的智能体会话**——**不**继承之前的会话 ID。手动重跑意味着你已经判定上一次的产出不行，再继续之前的对话只会重放被污染的上下文。（自动重试则相反，会继承会话——那条路径处理的是基础设施层面的失败，不是产出不好。）
 
 对比：
 
@@ -85,8 +88,9 @@ multica issue rerun <issue-id>
 |---|---|---|
 | 触发 | 系统基于失败原因自动执行 | 你主动发起 |
 | 上限 | 2 次 | 无上限 |
-| 适用来源 | issue、聊天 | 所有来源 |
-| 会话继承 | 是 | 是 |
+| 适用来源 | issue、聊天 | 有智能体分配人的 issue |
+| 跑哪个 agent | 失败任务原本的 agent | issue 当前的分配人 |
+| 会话继承 | 是（接着上次会话） | 否（全新会话） |
 
 ## 失败的任务对 issue 状态有什么影响
 
@@ -96,7 +100,7 @@ multica issue rerun <issue-id>
 
 可以——前提是对应的 AI 编程工具支持会话恢复。
 
-Multica 在任务过程中**两次**保存会话 ID——任务一开始（AI 工具返回第一条系统消息时）pin 一次，任务结束（完成或失败）时再 pin 一次。前者让守护进程中途崩溃时也能恢复，后者给之后的重跑用。下次重跑或自动重试时把这个 ID 传回去，智能体就能接着上次的对话、文件状态继续。
+Multica 在任务过程中**两次**保存会话 ID——任务一开始（AI 工具返回第一条系统消息时）pin 一次，任务结束（完成或失败）时再 pin 一次。前者让守护进程中途崩溃时也能恢复，后者留给下一次**自动重试**——届时把这个 ID 传回去，智能体就能接着上次的对话和文件状态继续。**手动重跑会主动跳过这一步**，永远从全新会话开始——见 [手动重跑和自动重试的区别](#手动重跑和自动重试的区别)。
 
 但**哪些 AI 编程工具真的支持**差别很大：
 

apps/docs/lib/locale-link.test.ts

@@ -0,0 +1,52 @@
+import { describe, expect, it } from "vitest";
+import { prefixLocale } from "./locale-link";
+
+describe("prefixLocale", () => {
+  it("prefixes root-relative paths with the active non-default locale", () => {
+    expect(prefixLocale("/workspaces", "zh")).toBe("/zh/workspaces");
+    expect(prefixLocale("/agents-create", "zh")).toBe("/zh/agents-create");
+  });
+
+  it("preserves anchors and query strings on prefixed paths", () => {
+    expect(prefixLocale("/providers#claude-code", "zh")).toBe(
+      "/zh/providers#claude-code",
+    );
+    expect(prefixLocale("/agents?from=docs", "zh")).toBe(
+      "/zh/agents?from=docs",
+    );
+  });
+
+  it("rewrites the bare root path to the locale root", () => {
+    expect(prefixLocale("/", "zh")).toBe("/zh");
+  });
+
+  it("leaves the default language untouched (URLs are prefix-less)", () => {
+    expect(prefixLocale("/workspaces", "en")).toBe("/workspaces");
+    expect(prefixLocale("/", "en")).toBe("/");
+  });
+
+  it("does not double-prefix paths that already carry a known locale", () => {
+    expect(prefixLocale("/zh/workspaces", "zh")).toBe("/zh/workspaces");
+    expect(prefixLocale("/en/workspaces", "zh")).toBe("/en/workspaces");
+  });
+
+  it("leaves external URLs alone", () => {
+    expect(prefixLocale("https://multica.ai/download", "zh")).toBe(
+      "https://multica.ai/download",
+    );
+    expect(prefixLocale("mailto:hello@multica.ai", "zh")).toBe(
+      "mailto:hello@multica.ai",
+    );
+    expect(prefixLocale("tel:+1234567890", "zh")).toBe("tel:+1234567890");
+  });
+
+  it("leaves in-page anchors and relative paths alone", () => {
+    expect(prefixLocale("#section", "zh")).toBe("#section");
+    expect(prefixLocale("./sibling", "zh")).toBe("./sibling");
+    expect(prefixLocale("../sibling", "zh")).toBe("../sibling");
+  });
+
+  it("returns empty/undefined hrefs unchanged", () => {
+    expect(prefixLocale("", "zh")).toBe("");
+  });
+});

apps/docs/lib/locale-link.ts

@@ -0,0 +1,31 @@
+import { i18n } from "./i18n";
+
+// Add the active locale prefix to root-relative MDX links so internal
+// navigation inside Chinese (or any non-default-language) docs stays in
+// that language. Without this, `[xx](/workspaces)` written in a `*.zh.mdx`
+// renders as `<a href="/workspaces">`, which Next's basePath rewrites to
+// `/docs/workspaces` and the docs middleware then routes to English —
+// leaking the reader out of their chosen locale.
+//
+// We deliberately do NOT touch:
+//   - external links (`https:`, `mailto:`, `tel:`, etc.)
+//   - in-page anchors (`#section`)
+//   - relative paths (`./foo`, `../bar`)
+//   - paths already prefixed with a known locale
+//   - the default language (URLs are intentionally prefix-less under
+//     `hideLocale: 'default-locale'`)
+export function prefixLocale(href: string, lang: string): string {
+  if (!href) return href;
+  if (lang === i18n.defaultLanguage) return href;
+  if (/^[a-z][a-z0-9+.-]*:/i.test(href)) return href;
+  if (href.startsWith("#")) return href;
+  if (!href.startsWith("/")) return href;
+
+  const segments = href.split("/").filter(Boolean);
+  const first = segments[0];
+  if (first && (i18n.languages as readonly string[]).includes(first)) {
+    return href;
+  }
+
+  return href === "/" ? `/${lang}` : `/${lang}${href}`;
+}

apps/docs/package.json

@@ -8,6 +8,7 @@
     "build": "fumadocs-mdx && next build",
     "start": "next start",
     "typecheck": "fumadocs-mdx && tsc --noEmit",
+    "test": "vitest run",
     "postinstall": "fumadocs-mdx"
   },
   "dependencies": {
@@ -27,6 +28,7 @@
     "@types/react": "catalog:",
     "@types/react-dom": "catalog:",
     "tailwindcss": "catalog:",
-    "typescript": "catalog:"
+    "typescript": "catalog:",
+    "vitest": "catalog:"
   }
 }

apps/docs/vitest.config.ts

@@ -0,0 +1,16 @@
+import { defineConfig } from "vitest/config";
+import path from "path";
+
+export default defineConfig({
+  test: {
+    environment: "node",
+    globals: true,
+    include: ["**/*.test.{ts,tsx}"],
+    exclude: ["node_modules/**", ".next/**", ".source/**"],
+  },
+  resolve: {
+    alias: {
+      "@": path.resolve(__dirname, "."),
+    },
+  },
+});

apps/web/app/(auth)/invitations/page.tsx

@@ -0,0 +1,28 @@
+"use client";
+
+import { useEffect } from "react";
+import { useRouter } from "next/navigation";
+import { useAuthStore } from "@multica/core/auth";
+import { paths } from "@multica/core/paths";
+import { InvitationsPage } from "@multica/views/invitations";
+
+export default function InvitationsRoutePage() {
+  const router = useRouter();
+  const user = useAuthStore((s) => s.user);
+  const isLoading = useAuthStore((s) => s.isLoading);
+
+  // Unauthenticated users have nowhere meaningful to land here — kick them
+  // through login and bring them back. The login page will eventually run
+  // its own listMyInvitations() check and route them here again.
+  useEffect(() => {
+    if (!isLoading && !user) {
+      router.replace(
+        `${paths.login()}?next=${encodeURIComponent(paths.invitations())}`,
+      );
+    }
+  }, [isLoading, user, router]);
+
+  if (isLoading || !user) return null;
+
+  return <InvitationsPage />;
+}

apps/web/app/(auth)/login/page.test.tsx

@@ -2,12 +2,22 @@ import { describe, it, expect, vi, beforeEach } from "vitest";
 import { render, screen, waitFor } from "@testing-library/react";
 import userEvent from "@testing-library/user-event";
 import { QueryClient, QueryClientProvider } from "@tanstack/react-query";
+import { I18nProvider } from "@multica/core/i18n/react";
+import enCommon from "@multica/views/locales/en/common.json";
+import enAuth from "@multica/views/locales/en/auth.json";
+import enSettings from "@multica/views/locales/en/settings.json";
 import type { ReactNode } from "react";
 
+const TEST_RESOURCES = {
+  en: { common: enCommon, auth: enAuth, settings: enSettings },
+};
+
 function createWrapper() {
   const qc = new QueryClient({ defaultOptions: { queries: { retry: false } } });
   return ({ children }: { children: ReactNode }) => (
-    <QueryClientProvider client={qc}>{children}</QueryClientProvider>
+    <I18nProvider locale="en" resources={TEST_RESOURCES}>
+      <QueryClientProvider client={qc}>{children}</QueryClientProvider>
+    </I18nProvider>
   );
 }
 

apps/web/app/(auth)/login/page.tsx

@@ -2,7 +2,7 @@
 
 import { Suspense, useEffect, useState } from "react";
 import { useSearchParams, useRouter } from "next/navigation";
-import { useQueryClient } from "@tanstack/react-query";
+import { useQueryClient, type QueryClient } from "@tanstack/react-query";
 import { sanitizeNextUrl, useAuthStore } from "@multica/core/auth";
 import { useConfigStore } from "@multica/core/config";
 import { workspaceKeys } from "@multica/core/workspace/queries";
@@ -26,10 +26,38 @@ import { captureDownloadIntent } from "@multica/core/analytics";
 import { setLoggedInCookie } from "@/features/auth/auth-cookie";
 import Link from "next/link";
 import { LoginPage, validateCliCallback } from "@multica/views/auth";
+import { useT } from "@multica/views/i18n";
+
+/**
+ * Pick where a logged-in user with no explicit `?next=` should land.
+ * Un-onboarded users with pending invitations on their email get routed to
+ * the batch /invitations page; everyone else falls through to the standard
+ * resolver. A network blip on listMyInvitations is non-fatal — we fall
+ * through rather than trap the user on an error screen.
+ */
+async function resolveLoggedInDestination(
+  qc: QueryClient,
+  hasOnboarded: boolean,
+  workspaces: Workspace[],
+): Promise<string> {
+  if (!hasOnboarded) {
+    try {
+      const invites = await api.listMyInvitations();
+      if (invites.length > 0) {
+        qc.setQueryData(workspaceKeys.myInvitations(), invites);
+        return paths.invitations();
+      }
+    } catch {
+      // fall through
+    }
+  }
+  return resolvePostAuthDestination(workspaces, hasOnboarded);
+}
 
 function LoginPageContent() {
   const router = useRouter();
   const qc = useQueryClient();
+  const { t } = useT("auth");
   const googleClientId = useConfigStore((state) => state.googleClientId);
   const user = useAuthStore((s) => s.user);
   const isLoading = useAuthStore((s) => s.isLoading);
@@ -67,38 +95,35 @@ function LoginPageContent() {
         })
         .catch((err) => {
           setDesktopError(
-            err instanceof Error ? err.message : "Failed to prepare Desktop sign-in",
+            err instanceof Error
+              ? err.message
+              : t(($) => $.web.desktop_handoff.prepare_failed),
           );
         });
       return;
     }
-    if (!hasOnboarded) {
-      router.replace(paths.onboarding());
-      return;
-    }
     if (nextUrl) {
       router.replace(nextUrl);
       return;
     }
     const list = qc.getQueryData<Workspace[]>(workspaceKeys.list()) ?? [];
-    router.replace(resolvePostAuthDestination(list, hasOnboarded));
+    void resolveLoggedInDestination(qc, hasOnboarded, list).then((dest) =>
+      router.replace(dest),
+    );
   }, [isLoading, user, router, nextUrl, cliCallbackRaw, isDesktopHandoff, hasOnboarded, qc]);
 
-  const handleSuccess = () => {
+  const handleSuccess = async () => {
     // Read the latest user snapshot directly — the closure's `hasOnboarded`
     // was captured before login completed and would be stale here.
     const currentUser = useAuthStore.getState().user;
     const onboarded = currentUser?.onboarded_at != null;
-    if (!onboarded) {
-      router.push(paths.onboarding());
-      return;
-    }
     if (nextUrl) {
       router.push(nextUrl);
       return;
     }
     const list = qc.getQueryData<Workspace[]>(workspaceKeys.list()) ?? [];
-    router.push(resolvePostAuthDestination(list, onboarded));
+    const dest = await resolveLoggedInDestination(qc, onboarded, list);
+    router.push(dest);
   };
 
   // Build Google OAuth state: encode platform + next URL so the callback
@@ -119,7 +144,9 @@ function LoginPageContent() {
         <div className="flex min-h-screen items-center justify-center">
           <Card className="w-full max-w-sm">
             <CardHeader className="text-center">
-              <CardTitle className="text-2xl">Sign-in Failed</CardTitle>
+              <CardTitle className="text-2xl">
+                {t(($) => $.web.desktop_handoff.failed_title)}
+              </CardTitle>
               <CardDescription>{desktopError}</CardDescription>
             </CardHeader>
           </Card>
@@ -130,11 +157,13 @@ function LoginPageContent() {
       <div className="flex min-h-screen items-center justify-center">
         <Card className="w-full max-w-sm">
           <CardHeader className="text-center">
-            <CardTitle className="text-2xl">Opening Multica</CardTitle>
+            <CardTitle className="text-2xl">
+              {t(($) => $.web.desktop_handoff.opening_title)}
+            </CardTitle>
             <CardDescription>
               {desktopToken
-                ? "You should see a prompt to open the Multica desktop app. If nothing happens, click the button below."
-                : "Preparing Desktop sign-in..."}
+                ? t(($) => $.web.desktop_handoff.opening_description)
+                : t(($) => $.web.desktop_handoff.preparing)}
             </CardDescription>
           </CardHeader>
           <CardContent className="flex justify-center">
@@ -145,7 +174,7 @@ function LoginPageContent() {
                   window.location.href = `multica://auth/callback?token=${encodeURIComponent(desktopToken)}`;
                 }}
               >
-                Open Multica Desktop
+                {t(($) => $.web.desktop_handoff.open_button)}
               </Button>
             ) : (
               <Loader2 className="h-6 w-6 animate-spin text-muted-foreground" />
@@ -175,18 +204,14 @@ function LoginPageContent() {
       }
       onTokenObtained={setLoggedInCookie}
       extra={
-        // Web-only nudge toward the desktop app. Copy is hardcoded EN
-        // for now because the login route sits outside the landing
-        // group's LocaleProvider — if this page ever becomes
-        // locale-aware, the strings live in positioning doc §3.3.
         <span className="text-xs text-muted-foreground">
-          Prefer the desktop app?{" "}
+          {t(($) => $.web.prefer_desktop)}{" "}
           <Link
             href="/download"
             onClick={() => captureDownloadIntent("login")}
             className="font-medium text-foreground underline decoration-foreground/30 underline-offset-4 hover:decoration-foreground/70"
           >
-            Download
+            {t(($) => $.web.download)}
           </Link>
         </span>
       }

apps/web/app/(auth)/onboarding/page.tsx

@@ -32,7 +32,7 @@ export default function OnboardingPage() {
   const hasOnboarded = useHasOnboarded();
   const { data: workspaces = [], isFetched: workspacesFetched } = useQuery({
     ...workspaceListOptions(),
-    enabled: !!user && hasOnboarded,
+    enabled: !!user,
   });
 
   useEffect(() => {
@@ -40,7 +40,15 @@ export default function OnboardingPage() {
       if (!isLoading && !user) router.replace(paths.login());
       return;
     }
-    if (hasOnboarded && workspacesFetched) {
+    if (!workspacesFetched) return;
+    // Bounce out only when onboarding genuinely doesn't apply: the user is
+    // already onboarded. We deliberately don't bounce on `workspaces.length`
+    // here — Step 3 of the flow creates a workspace mid-onboarding, and a
+    // hasWorkspaces bounce here would kick the user out before Steps 4–5
+    // (runtime / agent / first issue) can run. The new entry-point
+    // judgment in callback / login handles "where should this user go on
+    // login" so OnboardingPage no longer needs to second-guess it.
+    if (hasOnboarded) {
       router.replace(resolvePostAuthDestination(workspaces, hasOnboarded));
     }
   }, [isLoading, user, hasOnboarded, workspacesFetched, workspaces, router]);

apps/web/app/(landing)/layout.tsx

@@ -1,5 +1,6 @@
 import { cookies, headers } from "next/headers";
 import { Instrument_Serif, Noto_Serif_SC } from "next/font/google";
+import { LOCALE_COOKIE } from "@multica/core/i18n";
 import { LocaleProvider } from "@/features/landing/i18n";
 import type { Locale } from "@/features/landing/i18n";
 
@@ -43,7 +44,7 @@ const jsonLd = {
 async function getInitialLocale(): Promise<Locale> {
   // 1. User's explicit preference (cookie set when they switch language)
   const cookieStore = await cookies();
-  const stored = cookieStore.get("multica-locale")?.value;
+  const stored = cookieStore.get(LOCALE_COOKIE)?.value;
   if (stored === "en" || stored === "zh") return stored;
 
   // 2. Detect from Accept-Language header

apps/web/app/[workspaceSlug]/(dashboard)/issues/[id]/page.tsx

@@ -2,6 +2,7 @@
 
 import { use } from "react";
 import { IssueDetail } from "@multica/views/issues/components";
+import { ErrorBoundary } from "@multica/ui/components/common/error-boundary";
 
 export default function IssueDetailPage({
   params,
@@ -9,5 +10,9 @@ export default function IssueDetailPage({
   params: Promise<{ id: string }>;
 }) {
   const { id } = use(params);
-  return <IssueDetail issueId={id} />;
+  return (
+    <ErrorBoundary resetKeys={[id]}>
+      <IssueDetail issueId={id} />
+    </ErrorBoundary>
+  );
 }

apps/web/app/[workspaceSlug]/(dashboard)/issues/page.tsx

@@ -1,7 +1,12 @@
 "use client";
 
 import { IssuesPage } from "@multica/views/issues/components";
+import { ErrorBoundary } from "@multica/ui/components/common/error-boundary";
 
 export default function Page() {
-  return <IssuesPage />;
+  return (
+    <ErrorBoundary>
+      <IssuesPage />
+    </ErrorBoundary>
+  );
 }

apps/web/app/auth/callback/page.test.tsx

@@ -2,13 +2,21 @@ import { describe, it, expect, vi, beforeEach } from "vitest";
 import { render, waitFor } from "@testing-library/react";
 import { paths } from "@multica/core/paths";
 
-const { mockPush, mockSearchParams, mockLoginWithGoogle, mockListWorkspaces } =
-  vi.hoisted(() => ({
-    mockPush: vi.fn(),
-    mockSearchParams: new URLSearchParams(),
-    mockLoginWithGoogle: vi.fn(),
-    mockListWorkspaces: vi.fn(),
-  }));
+const {
+  mockPush,
+  mockSearchParams,
+  mockLoginWithGoogle,
+  mockListWorkspaces,
+  mockListMyInvitations,
+  mockSetQueryData,
+} = vi.hoisted(() => ({
+  mockPush: vi.fn(),
+  mockSearchParams: new URLSearchParams(),
+  mockLoginWithGoogle: vi.fn(),
+  mockListWorkspaces: vi.fn(),
+  mockListMyInvitations: vi.fn(),
+  mockSetQueryData: vi.fn(),
+}));
 
 const makeUser = (overrides: Partial<{ onboarded_at: string | null }> = {}) => ({
   id: "user-1",
@@ -28,7 +36,7 @@ vi.mock("next/navigation", () => ({
 }));
 
 vi.mock("@tanstack/react-query", () => ({
-  useQueryClient: () => ({ setQueryData: vi.fn() }),
+  useQueryClient: () => ({ setQueryData: mockSetQueryData }),
 }));
 
 // Preserve the real sanitizeNextUrl so the "drop unsafe ?next=" behavior is
@@ -46,12 +54,16 @@ vi.mock("@multica/core/auth", async () => {
 });
 
 vi.mock("@multica/core/workspace/queries", () => ({
-  workspaceKeys: { list: () => ["workspaces"] },
+  workspaceKeys: {
+    list: () => ["workspaces"],
+    myInvitations: () => ["invitations", "mine"],
+  },
 }));
 
 vi.mock("@multica/core/api", () => ({
   api: {
     listWorkspaces: mockListWorkspaces,
+    listMyInvitations: mockListMyInvitations,
     googleLogin: vi.fn(),
   },
 }));
@@ -61,26 +73,78 @@ import CallbackPage from "./page";
 describe("CallbackPage", () => {
   beforeEach(() => {
     vi.clearAllMocks();
-    mockSearchParams.forEach((_v, k) => mockSearchParams.delete(k));
+    // Snapshot keys before deleting — forEach + delete skips entries because
+    // the iteration index advances while the underlying list shrinks.
+    Array.from(mockSearchParams.keys()).forEach((k) =>
+      mockSearchParams.delete(k),
+    );
     mockSearchParams.set("code", "test-code");
     mockLoginWithGoogle.mockResolvedValue(makeUser());
     mockListWorkspaces.mockResolvedValue([]);
+    mockListMyInvitations.mockResolvedValue([]);
   });
 
-  it("unonboarded user lands on /onboarding regardless of next=", async () => {
+  it("unonboarded user honors a safe next= (e.g. /invite/{id}) so invitees aren't trapped", async () => {
     mockSearchParams.set("state", "next:/invite/abc123");
     render(<CallbackPage />);
     await waitFor(() => {
-      expect(mockPush).toHaveBeenCalledWith(paths.onboarding());
+      expect(mockPush).toHaveBeenCalledWith("/invite/abc123");
     });
-    expect(mockPush).not.toHaveBeenCalledWith("/invite/abc123");
+    expect(mockPush).not.toHaveBeenCalledWith(paths.onboarding());
+    // nextUrl is a fast path — listMyInvitations should not be queried.
+    expect(mockListMyInvitations).not.toHaveBeenCalled();
   });
 
-  it("unonboarded user with no next= also lands on /onboarding", async () => {
+  it("unonboarded user with no next= and no pending invitations lands on /onboarding", async () => {
     render(<CallbackPage />);
     await waitFor(() => {
       expect(mockPush).toHaveBeenCalledWith(paths.onboarding());
     });
+    expect(mockListMyInvitations).toHaveBeenCalled();
+  });
+
+  it("unonboarded user with pending invitations lands on /invitations", async () => {
+    mockListMyInvitations.mockResolvedValue([
+      {
+        id: "inv-1",
+        workspace_id: "ws-1",
+        workspace_name: "Acme",
+        role: "member",
+        status: "pending",
+      },
+    ]);
+    render(<CallbackPage />);
+    await waitFor(() => {
+      expect(mockPush).toHaveBeenCalledWith(paths.invitations());
+    });
+    expect(mockPush).not.toHaveBeenCalledWith(paths.onboarding());
+  });
+
+  it("onboarded user with workspace lands in that workspace", async () => {
+    mockLoginWithGoogle.mockResolvedValue(
+      makeUser({ onboarded_at: "2026-01-01T00:00:00Z" }),
+    );
+    mockListWorkspaces.mockResolvedValue([
+      {
+        id: "ws-1",
+        name: "Acme",
+        slug: "acme",
+        description: null,
+        context: null,
+        settings: {},
+        repos: [],
+        issue_prefix: "ACME",
+        created_at: "",
+        updated_at: "",
+      },
+    ]);
+    render(<CallbackPage />);
+    await waitFor(() => {
+      expect(mockPush).toHaveBeenCalledWith(paths.workspace("acme").issues());
+    });
+    // Already-onboarded users skip the listMyInvitations check; new invites
+    // surface in the sidebar instead of the wall.
+    expect(mockListMyInvitations).not.toHaveBeenCalled();
   });
 
   it("onboarded user ignores unsafe next= targets and lands on the default destination", async () => {
@@ -109,4 +173,12 @@ describe("CallbackPage", () => {
       expect(mockPush).toHaveBeenCalledWith("/invite/abc123");
     });
   });
+
+  it("falls through to /onboarding when listMyInvitations errors", async () => {
+    mockListMyInvitations.mockRejectedValue(new Error("network"));
+    render(<CallbackPage />);
+    await waitFor(() => {
+      expect(mockPush).toHaveBeenCalledWith(paths.onboarding());
+    });
+  });
 });

apps/web/app/auth/callback/page.tsx

@@ -66,13 +66,42 @@ function CallbackContent() {
           const wsList = await api.listWorkspaces();
           qc.setQueryData(workspaceKeys.list(), wsList);
           const onboarded = loggedInUser.onboarded_at != null;
-          if (!onboarded) {
-            router.push(paths.onboarding());
+
+          // 1. nextUrl wins: a `next=/invite/<id>` always survives the OAuth
+          //    round-trip — the user clicked a specific link and we should
+          //    honor exactly that destination.
+          if (nextUrl) {
+            router.push(nextUrl);
             return;
           }
-          router.push(
-            nextUrl || resolvePostAuthDestination(wsList, onboarded),
-          );
+
+          // 2. Un-onboarded users may have pending invitations on their
+          //    email even when no `next=` was carried (came from a fresh
+          //    login on app.multica.ai instead of clicking the email link,
+          //    or `state` was lost across the round-trip). Look them up by
+          //    email and route to the batch /invitations page if any.
+          //    Already-onboarded users skip this lookup — their new invites
+          //    surface in the sidebar dropdown, not as a forced wall.
+          if (!onboarded) {
+            try {
+              const invites = await api.listMyInvitations();
+              if (invites.length > 0) {
+                qc.setQueryData(workspaceKeys.myInvitations(), invites);
+                router.push(paths.invitations());
+                return;
+              }
+            } catch {
+              // Network blip on the invite lookup is non-fatal — fall through
+              // to the normal post-auth destination so the user isn't stuck
+              // on a blank callback screen. Worst case they land on
+              // /onboarding and the sidebar will surface invites later.
+            }
+          }
+
+          // 3. Default: hand off to the resolver (onboarding for first-timers,
+          //    first workspace for returning users, /workspaces/new for
+          //    onboarded users with zero workspaces).
+          router.push(resolvePostAuthDestination(wsList, onboarded));
         })
         .catch((err) => {
           setError(err instanceof Error ? err.message : "Login failed");

apps/web/app/layout.tsx

@@ -1,10 +1,16 @@
 import type { Metadata, Viewport } from "next";
+import { headers } from "next/headers";
 import { Inter, Geist_Mono, Source_Serif_4 } from "next/font/google";
 import { ThemeProvider } from "@/components/theme-provider";
 import { Toaster } from "@multica/ui/components/ui/sonner";
 import { cn } from "@multica/ui/lib/utils";
 import { WebProviders } from "@/components/web-providers";
-import { LocaleSync } from "@/components/locale-sync";
+import {
+  DEFAULT_LOCALE,
+  SUPPORTED_LOCALES,
+  type SupportedLocale,
+} from "@multica/core/i18n";
+import { RESOURCES } from "@multica/views/locales";
 import "./globals.css";
 
 // Font stack: Inter for Latin UI text + system Chinese fonts for zh content.
@@ -97,21 +103,40 @@ export const metadata: Metadata = {
   },
 };
 
-export default function RootLayout({
+function isSupportedLocale(value: string | null): value is SupportedLocale {
+  return value !== null && (SUPPORTED_LOCALES as readonly string[]).includes(value);
+}
+
+// HTML lang attribute uses BCP-47 region tags that screen readers and font
+// stacks recognize widely. i18next keeps `zh-Hans` as its internal locale
+// (script subtag is what we actually translate against), but the html element
+// expects a region-flavoured tag for accessibility tooling and CJK fallback.
+const HTML_LANG: Record<SupportedLocale, string> = {
+  en: "en",
+  "zh-Hans": "zh-CN",
+};
+
+export default async function RootLayout({
   children,
 }: {
   children: React.ReactNode;
 }) {
+  const h = await headers();
+  const headerLocale = h.get("x-multica-locale");
+  const locale: SupportedLocale = isSupportedLocale(headerLocale)
+    ? headerLocale
+    : DEFAULT_LOCALE;
+  const resources = { [locale]: RESOURCES[locale] };
+
   return (
     <html
-      lang="en"
+      lang={HTML_LANG[locale]}
       suppressHydrationWarning
       className={cn("antialiased font-sans h-full", inter.variable, geistMono.variable, sourceSerif.variable)}
     >
       <body className="h-full overflow-hidden">
-        <LocaleSync />
         <ThemeProvider>
-          <WebProviders>
+          <WebProviders locale={locale} resources={resources}>
             {children}
           </WebProviders>
           <Toaster />

apps/web/app/not-found.tsx

@@ -0,0 +1,19 @@
+"use client";
+
+import Link from "next/link";
+import { buttonVariants } from "@multica/ui/components/ui/button";
+
+export default function NotFound() {
+  return (
+    <main className="flex min-h-screen flex-col items-center justify-center gap-6 px-6 py-24 text-center">
+      <p className="text-sm font-medium text-muted-foreground">404</p>
+      <h1 className="text-2xl font-semibold tracking-tight">Page not found</h1>
+      <p className="max-w-md text-sm text-muted-foreground">
+        The page you are looking for doesn&rsquo;t exist or has been moved.
+      </p>
+      <Link href="/" className={buttonVariants({ className: "mt-2" })}>
+        Back to Multica
+      </Link>
+    </main>
+  );
+}

apps/web/components/locale-sync.tsx

@@ -1,20 +0,0 @@
-"use client";
-
-import { useEffect } from "react";
-
-/**
- * Reads the locale cookie on the client and updates <html lang>.
- * This avoids calling cookies() in the root Server Component layout,
- * which would mark the entire app as dynamic and disable the Router Cache.
- */
-export function LocaleSync() {
-  useEffect(() => {
-    const match = document.cookie.match(/(?:^|;\s*)multica-locale=(\w+)/);
-    const locale = match?.[1];
-    if (locale === "zh") {
-      document.documentElement.lang = "zh";
-    }
-  }, []);
-
-  return null;
-}

apps/web/components/web-providers.tsx

@@ -2,6 +2,8 @@
 
 import { Suspense, useMemo } from "react";
 import { CoreProvider } from "@multica/core/platform";
+import { createBrowserCookieLocaleAdapter } from "@multica/core/i18n/browser";
+import type { LocaleResources, SupportedLocale } from "@multica/core/i18n";
 import packageJson from "../package.json";
 import { WebNavigationProvider } from "@/platform/navigation";
 import {
@@ -41,7 +43,15 @@ function deriveWsUrl(): string | undefined {
 const WEB_VERSION =
   process.env.NEXT_PUBLIC_APP_VERSION || packageJson.version || "dev";
 
-export function WebProviders({ children }: { children: React.ReactNode }) {
+export function WebProviders({
+  children,
+  locale,
+  resources,
+}: {
+  children: React.ReactNode;
+  locale: SupportedLocale;
+  resources: Record<string, LocaleResources>;
+}) {
   const cookieAuth = !hasLegacyToken();
   // Stable identity reference so downstream effects keyed on it don't see a
   // new object on every parent render.
@@ -49,6 +59,7 @@ export function WebProviders({ children }: { children: React.ReactNode }) {
     () => ({ platform: "web", version: WEB_VERSION }),
     [],
   );
+  const localeAdapter = useMemo(() => createBrowserCookieLocaleAdapter(), []);
   return (
     <CoreProvider
       apiBaseUrl={process.env.NEXT_PUBLIC_API_URL}
@@ -57,6 +68,9 @@ export function WebProviders({ children }: { children: React.ReactNode }) {
       onLogin={setLoggedInCookie}
       onLogout={clearLoggedInCookie}
       identity={identity}
+      locale={locale}
+      resources={resources}
+      localeAdapter={localeAdapter}
     >
       {/* Suspense boundary is required by Next.js for useSearchParams in
           a client component mounted this high in the tree. */}

apps/web/features/landing/i18n/context.tsx

@@ -2,6 +2,7 @@
 
 import { createContext, useContext, useState, useCallback, useMemo } from "react";
 import { useConfigStore } from "@multica/core/config";
+import { LOCALE_COOKIE } from "@multica/core/i18n";
 import { createEnDict } from "./en";
 import { createZhDict } from "./zh";
 import type { LandingDict, Locale } from "./types";
@@ -11,7 +12,6 @@ const dictionaryFactories: Record<Locale, (allowSignup: boolean) => LandingDict>
   zh: createZhDict,
 };
 
-const COOKIE_NAME = "multica-locale";
 const COOKIE_MAX_AGE = 60 * 60 * 24 * 365; // 1 year
 
 type LocaleContextValue = {
@@ -38,7 +38,11 @@ export function LocaleProvider({
 
   const setLocale = useCallback((l: Locale) => {
     setLocaleState(l);
-    document.cookie = `${COOKIE_NAME}=${l}; path=/; max-age=${COOKIE_MAX_AGE}; SameSite=Lax`;
+    const secure =
+      typeof location !== "undefined" && location.protocol === "https:"
+        ? "; Secure"
+        : "";
+    document.cookie = `${LOCALE_COOKIE}=${l}; path=/; max-age=${COOKIE_MAX_AGE}; SameSite=Lax${secure}`;
   }, []);
 
   return (

apps/web/features/landing/i18n/en.ts

@@ -94,7 +94,7 @@ export function createEnDict(allowSignup: boolean): LandingDict {
       label: "RUNTIMES",
       title: "One dashboard for all your compute",
       description:
-        "Local daemons and cloud runtimes, managed from a single panel. Real-time monitoring of online/offline status, usage charts, and activity heatmaps. Auto-detects local CLIs \u2014 plug in and go.",
+        "Local daemons and cloud runtimes, managed from a single panel. Real-time monitoring of online/offline status, usage charts, and activity heatmaps. Auto-detects 11 supported coding tools on your machine.",
       cards: [
         {
           title: "Unified runtime panel",
@@ -107,9 +107,9 @@ export function createEnDict(allowSignup: boolean): LandingDict {
             "Online/offline status, usage charts, and activity heatmaps. Know exactly what your compute is doing at any moment.",
         },
         {
-          title: "Auto-detection & plug-and-play",
+          title: "Auto-detection on first run",
           description:
-            "Multica detects available CLIs like Claude Code, Codex, OpenClaw, and OpenCode automatically. Connect a machine, and it\u2019s ready to work.",
+            "Multica scans for 11 supported coding tools \u2014 Claude Code, Codex, Cursor, Copilot, Gemini, Hermes, Kimi, Kiro CLI, OpenCode, OpenClaw, and Pi \u2014 and registers a runtime for each one it finds.",
         },
       ],
     },
@@ -129,7 +129,7 @@ export function createEnDict(allowSignup: boolean): LandingDict {
       {
         title: "Install the CLI & connect your machine",
         description:
-          "Run multica setup to configure, authenticate, and start the daemon. It auto-detects Claude Code, Codex, OpenClaw, and OpenCode on your machine \u2014 plug in and go.",
+          "Run multica setup \u2014 it walks you through OAuth, starts the daemon, and scans for the 11 supported coding tools (Claude Code, Codex, Cursor, Copilot, Gemini, Hermes, Kimi, Kiro CLI, OpenCode, OpenClaw, Pi). Whichever ones you already have installed get registered as runtimes automatically.",
       },
       {
         title: "Create your first agent",
@@ -185,7 +185,7 @@ export function createEnDict(allowSignup: boolean): LandingDict {
       {
         question: "What coding agents does Multica support?",
         answer:
-          "Multica currently supports Claude Code, Codex, OpenClaw, and OpenCode out of the box. The daemon auto-detects whichever CLIs you have installed. Since it\u2019s open source, you can also add your own backends.",
+          "Multica supports 11 coding tools out of the box: Claude Code, Codex, Cursor, Copilot, Gemini, Hermes, Kimi, Kiro CLI, OpenCode, OpenClaw, and Pi. The daemon auto-detects whichever CLIs you already have installed and registers a runtime for each one. Since it's open source, you can also add your own backends.",
       },
       {
         question: "Do I need to self-host, or is there a cloud version?",
@@ -283,6 +283,165 @@ export function createEnDict(allowSignup: boolean): LandingDict {
       fixes: "Bug Fixes",
     },
     entries: [
+      {
+        version: "0.2.28",
+        date: "2026-05-08",
+        title: "Daemon Disk-Usage CLI, Timeline Polish & Task Usage Rollup",
+        changes: [],
+        features: [
+          "New `multica daemon disk-usage` CLI surfaces per-task and per-workspace disk footprint",
+          "Skill picker in agent settings has a search box for fast lookup",
+          "Daemon GC extends to chat, autopilot, and quick-create tasks",
+          "Issue detail breadcrumb now shows the MUL-xxxx identifier for quick reference",
+        ],
+        improvements: [
+          "Timeline page size bumped to 50, with per-pool keyset cursors for comments and activities",
+          "'Show older / newer' affordances now appear in edge cases and look clearly clickable",
+          "Server `task_usage` rolls up into a daily aggregate table, dropping DB load significantly",
+          "Daemon health check stays responsive while repo lookups are in flight",
+          "Runtime stats exclude archived agents for accurate active counts",
+        ],
+        fixes: [
+          "Linux daemon self-restart uses `brew prefix` symlinks, so Homebrew Cellar deletion no longer orphans runtimes",
+          "CLI short IDs now route correctly — copied prefixes no longer 404",
+          "Windows non-ASCII comment / description input lands via new `--content-file` / `--description-file` flags",
+          "Windows / Linux desktop replaces the Electron placeholder icon with the Multica asterisk",
+          "Orphaned timeline replies are now correctly surfaced",
+          "Timeline comment pagination budget excludes activities, so heavy activity no longer crowds out real comments",
+        ],
+      },
+      {
+        version: "0.2.27",
+        date: "2026-05-07",
+        title: "Smoother Chat, GitHub Skill Import & Stability Fixes",
+        changes: [],
+        features: [
+          "Import reusable skills directly from GitHub links",
+        ],
+        improvements: [
+          "Chat and Inbox feel smoother, with clearer history, easier reply copying, and faster triage after archiving",
+          "Issue actions keep more context, from easier access to the local folder to sub-issues inheriting the right project and status",
+          "Autopilots pause themselves after repeated failures, so noisy automations are easier to catch and fix",
+        ],
+        fixes: [
+          "Chinese input, desktop updates, long issue timelines, and live status updates are more reliable",
+        ],
+      },
+      {
+        version: "0.2.26",
+        date: "2026-05-06",
+        title: "Full i18n Rollout, Long-Issue Timeline & System Notifications Toggle",
+        changes: [],
+        features: [
+          "Web app fully translated to Simplified Chinese (21 namespaces), with per-user locale",
+          "System Notifications toggle in Settings",
+          "Delete chat sessions; History panel surfaced on the chat header",
+          "Runtime liveness backed by Redis, with DB fallback",
+          "Desktop loads runtime self-host config",
+          "CLI adds `--assignee-id` / `--to-id` / `--user-id` for unambiguous targeting",
+        ],
+        improvements: [
+          "Settings 'Appearance' tab is renamed to 'Preferences', and the active tab is reflected in the URL so deep links work",
+          "Long issues open instantly — Timeline switched to cursor-based keyset pagination, and repeated `task_completed` / `task_failed` activity entries are coalesced",
+          "Runtime poll and heartbeat schedules are isolated per-runtime, so one busy runtime can no longer starve others",
+          "CLI update requests persist in Redis, so a server restart no longer drops them",
+          "Runtime cost usage window narrowed from 180 days to 14 days, dropping query load",
+          "Project list returns a `resource_count` instead of inlining all resources, keeping responses lean",
+          "404 page redesigned, with the No-Access redirect loop fixed",
+          "Quick Create exempts git-describe daemons from the CLI version gate",
+          "CI now enforces lint on every PR, and the existing lint debt has been cleared",
+        ],
+        fixes: [
+          "Daemon cancels the running agent when the task is deleted server-side, eliminating orphan processes",
+          "Daemon refreshes a stale Codex `auth.json` when reusing an exec env, fixing intermittent auth errors",
+          "Daemon refuses to write `.gc_meta.json` when `issue_id` is empty",
+          "Session / resume across ACP backends now trusts the agent-reported session id, fixing cross-session bleed",
+          "OpenCode skills are written under `.opencode/skills/` so they are discovered natively",
+          "404 task-not-found semantics tightened on both server and the final guard",
+          "Pinned sidebar rows are auto-unpinned when the underlying entity disappears",
+          "Project detail page splits desktop and mobile sidebar state",
+          "Runtime detail page hides archived agents",
+          "Already-attached repos in Add Resource show a URL tooltip; empty project state has a New Issue button",
+          "S3 public URLs are region-qualified, fixing cross-region access",
+          "Windows installer parses version numbers and decodes checksums correctly",
+          "Quick Create submit button no longer shows a duplicate keyboard shortcut",
+        ],
+      },
+      {
+        version: "0.2.24",
+        date: "2026-05-03",
+        title: "Repo Checkout `--ref`, Hermes Replay Fix & Multi-Replica Model Picker",
+        changes: [],
+        features: [
+          "`multica repo checkout --ref` targets a branch, tag, or specific commit when pulling a repo into the workspace",
+          "`multica agent avatar` uploads an agent avatar straight from the CLI",
+          "Inbox shows an archive button on done tasks; the redundant mark-as-done hover button is gone",
+        ],
+        improvements: [
+          "Long-timeline issues open instantly from Inbox — the markdown render pipeline is memoized so unrelated WS events no longer re-render thousands of comments",
+          "Model picker works on multi-replica deployments — pending requests persist via Redis, with daemon retries on transient report failures",
+          "Daemon empty-claim cache TTL bumped, further reducing idle DB load",
+        ],
+        fixes: [
+          "Newly created agents show up everywhere immediately — the agent cache is hydrated on create",
+          "Hermes no longer replays the previous answer when a new turn starts — historical chunks are gated behind a per-turn flag",
+          "Codex runtime model picker exposes the GPT-5.5 family",
+          "`multica login --token <PAT>` accepts the PAT as a flag value instead of rejecting it",
+          "CLI update completion status is now reliable",
+          "Session resume is guarded by runtime, preventing cross-runtime resume",
+          "Kanban display settings survive when dragging issues across columns",
+          "Autopilot list is responsive on mobile viewports",
+          "Quick Create prompts produce higher-fidelity descriptions from the user's input",
+          "Skill upsert sanitizes null bytes, fixing a PostgreSQL UTF8 error",
+          "Connect Remote dialog points to the correct install script URL",
+        ],
+      },
+      {
+        version: "0.2.21",
+        date: "2026-04-30",
+        title: "Quick Capture Overhaul, Mermaid Diagrams & Typed Project Resources",
+        changes: [],
+        features: [
+          "Quick Capture replaces the old New Issue dialog — continuous-create mode, file uploads, and automatic enrichment from pasted URLs",
+          "Mermaid diagrams render inline in markdown, with a fullscreen lightbox for complex graphs",
+          "Projects can bind their own repo, separate from the workspace default",
+          "Permission-aware UI across agents, comments, runtimes, and skills — actions you can't take are no longer offered",
+        ],
+        improvements: [
+          "Daemon `/tasks/claim` polling uses a Redis empty-claim fast-path, dropping idle DB load and reclaiming disk on long-open issues",
+          "Multica Agent commits include a `Co-authored-by` trailer for proper Git attribution",
+          "Desktop blocks Cmd+R / Ctrl+R / F5 from reloading the app and shows the real version in dev and Updates settings",
+        ],
+        fixes: [
+          "Quick Create no longer invents requirements beyond user input, and subscribes the requester to the issue it creates",
+          "Inbox jumps straight to the targeted comment, and auto-archives when the issue is marked Done from the detail page",
+          "Task rerun starts a fresh session and skips poisoned resume state",
+          "Invitees land on their workspace after sign-in instead of being forced through `/onboarding`",
+        ],
+      },
+      {
+        version: "0.2.20",
+        date: "2026-04-29",
+        title: "Create Issue by Agent, Agent Presence v3 & Daemon WebSocket Heartbeat",
+        changes: [],
+        features: [
+          "Create Issue by Agent — press `c`, write one line, pick an agent; issue creation runs async and the result lands in your inbox",
+          "Agent Presence v3 — availability and last-task split into clearer signals, with an execution log on the issue panel showing active and recent runs",
+          "Daemon ↔ server heartbeat now flows over WebSocket with HTTP fallback, cutting task wakeup latency",
+          "Mention picker ranks suggestions by your local recency",
+        ],
+        improvements: [
+          "Server caches PAT / daemon token lookups in Redis, so large fleets stop hammering the database on every request",
+          "Backend default agent CLI args via `MULTICA_CLAUDE_ARGS` / `MULTICA_CODEX_ARGS` env vars",
+          "Manual and agent create-issue flows share one dialog shell, and picker agents become the default assignee",
+        ],
+        fixes: [
+          "Create-issue-by-agent no longer leaves tasks stuck queued, and no longer duplicates the issue when an attachment upload fails",
+          "Agent comments respect newlines instead of rendering literal `\\n`, and multi-line replies keep their formatting",
+          "Agent-authored root comments no longer inherit parent @mentions, breaking accidental agent loops",
+          "Cursor agent on Windows preserves multi-line prompts",
+        ],
+      },
       {
         version: "0.2.19",
         date: "2026-04-28",

apps/web/features/landing/i18n/zh.ts

@@ -13,42 +13,42 @@ export function createZhDict(allowSignup: boolean): LandingDict {
     headlineLine1: "\u4f60\u7684\u4e0b\u4e00\u6279\u5458\u5de5",
     headlineLine2: "\u4e0d\u662f\u4eba\u7c7b\u3002",
     subheading:
-      "Multica \u662f\u4e00\u4e2a\u5f00\u6e90\u5e73\u53f0\uff0c\u5c06\u7f16\u7801 Agent \u53d8\u6210\u771f\u6b63\u7684\u961f\u53cb\u3002\u5206\u914d\u4efb\u52a1\u3001\u8ddf\u8e2a\u8fdb\u5ea6\u3001\u79ef\u7d2f\u6280\u80fd\u2014\u2014\u5728\u4e00\u4e2a\u5730\u65b9\u7ba1\u7406\u4f60\u7684\u4eba\u7c7b + Agent \u56e2\u961f\u3002",
+      "Multica \u662f\u4e00\u4e2a\u5f00\u6e90\u5e73\u53f0\uff0c\u5c06\u7f16\u7801 智能体 \u53d8\u6210\u771f\u6b63\u7684\u961f\u53cb\u3002\u5206\u914d\u4efb\u52a1\u3001\u8ddf\u8e2a\u8fdb\u5ea6\u3001\u79ef\u7d2f\u6280\u80fd\u2014\u2014\u5728\u4e00\u4e2a\u5730\u65b9\u7ba1\u7406\u4f60\u7684\u4eba\u7c7b + 智能体 \u56e2\u961f\u3002",
     cta: "免费开始",
     downloadDesktop: "下载桌面端",
     worksWith: "支持",
-    imageAlt: "Multica \u770b\u677f\u89c6\u56fe\u2014\u2014\u4eba\u7c7b\u548c Agent \u534f\u540c\u7ba1\u7406\u4efb\u52a1",
+    imageAlt: "Multica \u770b\u677f\u89c6\u56fe\u2014\u2014\u4eba\u7c7b\u548c 智能体 \u534f\u540c\u7ba1\u7406\u4efb\u52a1",
   },
 
   features: {
     teammates: {
       label: "\u56e2\u961f\u534f\u4f5c",
-      title: "\u50cf\u5206\u914d\u7ed9\u540c\u4e8b\u4e00\u6837\u5206\u914d\u7ed9 Agent",
+      title: "\u50cf\u5206\u914d\u7ed9\u540c\u4e8b\u4e00\u6837\u5206\u914d\u7ed9 智能体",
       description:
-        "Agent \u4e0d\u662f\u88ab\u52a8\u5de5\u5177\u2014\u2014\u5b83\u4eec\u662f\u4e3b\u52a8\u53c2\u4e0e\u8005\u3002\u5b83\u4eec\u62e5\u6709\u4e2a\u4eba\u8d44\u6599\u3001\u62a5\u544a\u72b6\u6001\u3001\u521b\u5efa Issue\u3001\u53d1\u8868\u8bc4\u8bba\u3001\u66f4\u65b0\u72b6\u6001\u3002\u4f60\u7684\u6d3b\u52a8\u6d41\u5c55\u793a\u4eba\u7c7b\u548c Agent \u5e76\u80a9\u5de5\u4f5c\u3002",
+        "智能体 \u4e0d\u662f\u88ab\u52a8\u5de5\u5177\u2014\u2014\u5b83\u4eec\u662f\u4e3b\u52a8\u53c2\u4e0e\u8005\u3002\u5b83\u4eec\u62e5\u6709\u4e2a\u4eba\u8d44\u6599\u3001\u62a5\u544a\u72b6\u6001\u3001\u521b\u5efa Issue\u3001\u53d1\u8868\u8bc4\u8bba\u3001\u66f4\u65b0\u72b6\u6001\u3002\u4f60\u7684\u6d3b\u52a8\u6d41\u5c55\u793a\u4eba\u7c7b\u548c 智能体 \u5e76\u80a9\u5de5\u4f5c\u3002",
       cards: [
         {
-          title: "Agent \u51fa\u73b0\u5728\u6307\u6d3e\u4eba\u9009\u62e9\u5668\u4e2d",
+          title: "智能体 \u51fa\u73b0\u5728\u6307\u6d3e\u4eba\u9009\u62e9\u5668\u4e2d",
           description:
-            "\u4eba\u7c7b\u548c Agent \u51fa\u73b0\u5728\u540c\u4e00\u4e2a\u4e0b\u62c9\u83dc\u5355\u91cc\u3002\u628a\u4efb\u52a1\u5206\u914d\u7ed9 Agent \u548c\u5206\u914d\u7ed9\u540c\u4e8b\u6ca1\u6709\u4efb\u4f55\u533a\u522b\u3002",
+            "\u4eba\u7c7b\u548c 智能体 \u51fa\u73b0\u5728\u540c\u4e00\u4e2a\u4e0b\u62c9\u83dc\u5355\u91cc\u3002\u628a\u4efb\u52a1\u5206\u914d\u7ed9 智能体 \u548c\u5206\u914d\u7ed9\u540c\u4e8b\u6ca1\u6709\u4efb\u4f55\u533a\u522b\u3002",
         },
         {
           title: "\u81ea\u4e3b\u53c2\u4e0e",
           description:
-            "Agent \u4e3b\u52a8\u521b\u5efa Issue\u3001\u53d1\u8868\u8bc4\u8bba\u3001\u66f4\u65b0\u72b6\u6001\u2014\u2014\u800c\u4e0d\u662f\u53ea\u5728\u88ab\u63d0\u793a\u65f6\u624d\u884c\u52a8\u3002",
+            "智能体 \u4e3b\u52a8\u521b\u5efa Issue\u3001\u53d1\u8868\u8bc4\u8bba\u3001\u66f4\u65b0\u72b6\u6001\u2014\u2014\u800c\u4e0d\u662f\u53ea\u5728\u88ab\u63d0\u793a\u65f6\u624d\u884c\u52a8\u3002",
         },
         {
           title: "\u7edf\u4e00\u7684\u6d3b\u52a8\u65f6\u95f4\u7ebf",
           description:
-            "\u6574\u4e2a\u56e2\u961f\u5171\u7528\u4e00\u4e2a\u6d3b\u52a8\u6d41\u3002\u4eba\u7c7b\u548c Agent \u7684\u64cd\u4f5c\u4ea4\u66ff\u5c55\u793a\uff0c\u4f60\u59cb\u7ec8\u77e5\u9053\u53d1\u751f\u4e86\u4ec0\u4e48\u3001\u662f\u8c01\u505a\u7684\u3002",
+            "\u6574\u4e2a\u56e2\u961f\u5171\u7528\u4e00\u4e2a\u6d3b\u52a8\u6d41\u3002\u4eba\u7c7b\u548c 智能体 \u7684\u64cd\u4f5c\u4ea4\u66ff\u5c55\u793a\uff0c\u4f60\u59cb\u7ec8\u77e5\u9053\u53d1\u751f\u4e86\u4ec0\u4e48\u3001\u662f\u8c01\u505a\u7684\u3002",
         },
       ],
     },
     autonomous: {
       label: "\u81ea\u4e3b\u6267\u884c",
-      title: "\u8bbe\u7f6e\u540e\u65e0\u9700\u7ba1\u7406\u2014\u2014Agent \u5728\u4f60\u7761\u89c9\u65f6\u5de5\u4f5c",
+      title: "\u8bbe\u7f6e\u540e\u65e0\u9700\u7ba1\u7406\u2014\u2014智能体 \u5728\u4f60\u7761\u89c9\u65f6\u5de5\u4f5c",
       description:
-        "\u4e0d\u53ea\u662f\u63d0\u793a-\u54cd\u5e94\u3002\u5b8c\u6574\u7684\u4efb\u52a1\u751f\u547d\u5468\u671f\u7ba1\u7406\uff1a\u5165\u961f\u3001\u9886\u53d6\u3001\u542f\u52a8\u3001\u5b8c\u6210\u6216\u5931\u8d25\u3002Agent \u4e3b\u52a8\u62a5\u544a\u963b\u585e\uff0c\u4f60\u901a\u8fc7 WebSocket \u83b7\u53d6\u5b9e\u65f6\u8fdb\u5ea6\u3002",
+        "\u4e0d\u53ea\u662f\u63d0\u793a-\u54cd\u5e94\u3002\u5b8c\u6574\u7684\u4efb\u52a1\u751f\u547d\u5468\u671f\u7ba1\u7406\uff1a\u5165\u961f\u3001\u9886\u53d6\u3001\u542f\u52a8\u3001\u5b8c\u6210\u6216\u5931\u8d25\u3002智能体 \u4e3b\u52a8\u62a5\u544a\u963b\u585e\uff0c\u4f60\u901a\u8fc7 WebSocket \u83b7\u53d6\u5b9e\u65f6\u8fdb\u5ea6\u3002",
       cards: [
         {
           title: "\u5b8c\u6574\u7684\u4efb\u52a1\u751f\u547d\u5468\u671f",
@@ -58,12 +58,12 @@ export function createZhDict(allowSignup: boolean): LandingDict {
         {
           title: "\u4e3b\u52a8\u62a5\u544a\u963b\u585e",
           description:
-            "\u5f53 Agent \u9047\u5230\u56f0\u96be\u65f6\uff0c\u4f1a\u7acb\u5373\u53d1\u51fa\u8b66\u62a5\u3002\u4e0d\u7528\u7b49\u51e0\u4e2a\u5c0f\u65f6\u540e\u624d\u53d1\u73b0\u4ec0\u4e48\u90fd\u6ca1\u53d1\u751f\u3002",
+            "\u5f53 智能体 \u9047\u5230\u56f0\u96be\u65f6\uff0c\u4f1a\u7acb\u5373\u53d1\u51fa\u8b66\u62a5\u3002\u4e0d\u7528\u7b49\u51e0\u4e2a\u5c0f\u65f6\u540e\u624d\u53d1\u73b0\u4ec0\u4e48\u90fd\u6ca1\u53d1\u751f\u3002",
         },
         {
           title: "\u5b9e\u65f6\u8fdb\u5ea6\u63a8\u9001",
           description:
-            "\u57fa\u4e8e WebSocket \u7684\u5b9e\u65f6\u66f4\u65b0\u3002\u5b9e\u65f6\u89c2\u770b Agent \u5de5\u4f5c\uff0c\u6216\u968f\u65f6\u67e5\u770b\u2014\u2014\u65f6\u95f4\u7ebf\u59cb\u7ec8\u662f\u6700\u65b0\u7684\u3002",
+            "\u57fa\u4e8e WebSocket \u7684\u5b9e\u65f6\u66f4\u65b0\u3002\u5b9e\u65f6\u89c2\u770b 智能体 \u5de5\u4f5c\uff0c\u6216\u968f\u65f6\u67e5\u770b\u2014\u2014\u65f6\u95f4\u7ebf\u59cb\u7ec8\u662f\u6700\u65b0\u7684\u3002",
         },
       ],
     },
@@ -71,22 +71,22 @@ export function createZhDict(allowSignup: boolean): LandingDict {
       label: "\u6280\u80fd\u5e93",
       title: "\u6bcf\u4e2a\u89e3\u51b3\u65b9\u6848\u90fd\u6210\u4e3a\u5168\u56e2\u961f\u53ef\u590d\u7528\u7684\u6280\u80fd",
       description:
-        "\u6280\u80fd\u662f\u53ef\u590d\u7528\u7684\u80fd\u529b\u5b9a\u4e49\u2014\u2014\u4ee3\u7801\u3001\u914d\u7f6e\u548c\u4e0a\u4e0b\u6587\u6253\u5305\u5728\u4e00\u8d77\u3002\u53ea\u9700\u7f16\u5199\u4e00\u6b21\uff0c\u56e2\u961f\u4e2d\u6bcf\u4e2a Agent \u90fd\u80fd\u4f7f\u7528\u3002\u4f60\u7684\u6280\u80fd\u5e93\u968f\u65f6\u95f4\u4e0d\u65ad\u79ef\u7d2f\u3002",
+        "\u6280\u80fd\u662f\u53ef\u590d\u7528\u7684\u80fd\u529b\u5b9a\u4e49\u2014\u2014\u4ee3\u7801\u3001\u914d\u7f6e\u548c\u4e0a\u4e0b\u6587\u6253\u5305\u5728\u4e00\u8d77\u3002\u53ea\u9700\u7f16\u5199\u4e00\u6b21\uff0c\u56e2\u961f\u4e2d\u6bcf\u4e2a 智能体 \u90fd\u80fd\u4f7f\u7528\u3002\u4f60\u7684\u6280\u80fd\u5e93\u968f\u65f6\u95f4\u4e0d\u65ad\u79ef\u7d2f\u3002",
       cards: [
         {
           title: "\u53ef\u590d\u7528\u7684\u6280\u80fd\u5b9a\u4e49",
           description:
-            "\u5c06\u77e5\u8bc6\u5c01\u88c5\u6210\u4efb\u4f55 Agent \u90fd\u80fd\u6267\u884c\u7684\u6280\u80fd\u3002\u90e8\u7f72\u5230\u6d4b\u8bd5\u73af\u5883\u3001\u7f16\u5199\u8fc1\u79fb\u3001\u5ba1\u67e5 PR\u2014\u2014\u5168\u90e8\u4ee3\u7801\u5316\u3002",
+            "\u5c06\u77e5\u8bc6\u5c01\u88c5\u6210\u4efb\u4f55 智能体 \u90fd\u80fd\u6267\u884c\u7684\u6280\u80fd\u3002\u90e8\u7f72\u5230\u6d4b\u8bd5\u73af\u5883\u3001\u7f16\u5199\u8fc1\u79fb\u3001\u5ba1\u67e5 PR\u2014\u2014\u5168\u90e8\u4ee3\u7801\u5316\u3002",
         },
         {
           title: "\u5168\u56e2\u961f\u5171\u4eab",
           description:
-            "\u4e00\u4e2a\u4eba\u7684\u6280\u80fd\u5c31\u662f\u6bcf\u4e2a Agent \u7684\u6280\u80fd\u3002\u7f16\u5199\u4e00\u6b21\uff0c\u5168\u56e2\u961f\u53d7\u76ca\u3002",
+            "\u4e00\u4e2a\u4eba\u7684\u6280\u80fd\u5c31\u662f\u6bcf\u4e2a 智能体 \u7684\u6280\u80fd\u3002\u7f16\u5199\u4e00\u6b21\uff0c\u5168\u56e2\u961f\u53d7\u76ca\u3002",
         },
         {
           title: "\u590d\u5408\u589e\u957f",
           description:
-            "\u7b2c 1 \u5929\uff1a\u4f60\u6559 Agent \u90e8\u7f72\u3002\u7b2c 30 \u5929\uff1a\u6bcf\u4e2a Agent \u90fd\u80fd\u90e8\u7f72\u3001\u5199\u6d4b\u8bd5\u3001\u505a\u4ee3\u7801\u5ba1\u67e5\u3002\u56e2\u961f\u80fd\u529b\u6307\u6570\u7ea7\u589e\u957f\u3002",
+            "\u7b2c 1 \u5929\uff1a\u4f60\u6559 智能体 \u90e8\u7f72\u3002\u7b2c 30 \u5929\uff1a\u6bcf\u4e2a 智能体 \u90fd\u80fd\u90e8\u7f72\u3001\u5199\u6d4b\u8bd5\u3001\u505a\u4ee3\u7801\u5ba1\u67e5\u3002\u56e2\u961f\u80fd\u529b\u6307\u6570\u7ea7\u589e\u957f\u3002",
         },
       ],
     },
@@ -94,7 +94,7 @@ export function createZhDict(allowSignup: boolean): LandingDict {
       label: "\u8fd0\u884c\u65f6",
       title: "\u4e00\u4e2a\u63a7\u5236\u53f0\u7ba1\u7406\u6240\u6709\u7b97\u529b",
       description:
-        "\u672c\u5730\u5b88\u62a4\u8fdb\u7a0b\u548c\u4e91\u7aef\u8fd0\u884c\u65f6\uff0c\u5728\u540c\u4e00\u4e2a\u9762\u677f\u4e2d\u7ba1\u7406\u3002\u5b9e\u65f6\u76d1\u63a7\u5728\u7ebf/\u79bb\u7ebf\u72b6\u6001\u3001\u4f7f\u7528\u91cf\u56fe\u8868\u548c\u6d3b\u52a8\u70ed\u529b\u56fe\u3002\u81ea\u52a8\u68c0\u6d4b\u672c\u5730 CLI\u2014\u2014\u63d2\u4e0a\u5c31\u7528\u3002",
+        "\u672c\u5730\u5b88\u62a4\u8fdb\u7a0b\u548c\u4e91\u7aef\u8fd0\u884c\u65f6\uff0c\u5728\u540c\u4e00\u4e2a\u9762\u677f\u4e2d\u7ba1\u7406\u3002\u5b9e\u65f6\u76d1\u63a7\u5728\u7ebf/\u79bb\u7ebf\u72b6\u6001\u3001\u4f7f\u7528\u91cf\u56fe\u8868\u548c\u6d3b\u52a8\u70ed\u529b\u56fe\u3002\u81ea\u52a8\u68c0\u6d4b\u672c\u673a\u5df2\u5b89\u88c5\u7684 11 \u6b3e\u652f\u6301\u7684 AI \u7f16\u7a0b\u5de5\u5177\u3002",
       cards: [
         {
           title: "\u7edf\u4e00\u8fd0\u884c\u65f6\u9762\u677f",
@@ -107,9 +107,9 @@ export function createZhDict(allowSignup: boolean): LandingDict {
             "\u5728\u7ebf/\u79bb\u7ebf\u72b6\u6001\u3001\u4f7f\u7528\u91cf\u56fe\u8868\u548c\u6d3b\u52a8\u70ed\u529b\u56fe\u3002\u968f\u65f6\u4e86\u89e3\u4f60\u7684\u7b97\u529b\u5728\u505a\u4ec0\u4e48\u3002",
         },
         {
-          title: "\u81ea\u52a8\u68c0\u6d4b\u4e0e\u5373\u63d2\u5373\u7528",
+          title: "\u9996\u6b21\u542f\u52a8\u81ea\u52a8\u6ce8\u518c",
           description:
-            "Multica \u81ea\u52a8\u68c0\u6d4b Claude Code\u3001Codex\u3001OpenClaw \u548c OpenCode \u7b49\u53ef\u7528 CLI\u3002\u8fde\u63a5\u4e00\u53f0\u673a\u5668\uff0c\u5373\u53ef\u5f00\u59cb\u5de5\u4f5c\u3002",
+            "Multica \u626b\u63cf\u672c\u673a\u7684 11 \u6b3e\u652f\u6301\u7684 AI \u7f16\u7a0b\u5de5\u5177\u2014\u2014Claude Code\u3001Codex\u3001Cursor\u3001Copilot\u3001Gemini\u3001Hermes\u3001Kimi\u3001Kiro CLI\u3001OpenCode\u3001OpenClaw\u3001Pi\u2014\u2014\u5e76\u4e3a\u6bcf\u6b3e\u5df2\u5b89\u88c5\u7684\u5de5\u5177\u6ce8\u518c\u4e00\u4e2a\u8fd0\u884c\u65f6\u3002",
         },
       ],
     },
@@ -129,17 +129,17 @@ export function createZhDict(allowSignup: boolean): LandingDict {
       {
         title: "\u5b89\u88c5 CLI \u5e76\u8fde\u63a5\u4f60\u7684\u673a\u5668",
         description:
-          "运行 multica setup 一键完成配置、认证和启动。守护进程自动检测你机器上的 Claude Code、Codex、OpenClaw 和 OpenCode——插上就用。",
+          "运行 multica setup——它会引导你完成 OAuth 登录、启动守护进程、并扫描 11 款支持的 AI 编程工具（Claude Code、Codex、Cursor、Copilot、Gemini、Hermes、Kimi、Kiro CLI、OpenCode、OpenClaw、Pi）。本机已安装的工具会被自动注册成运行时。",
       },
       {
-        title: "\u521b\u5efa\u4f60\u7684\u7b2c\u4e00\u4e2a Agent",
+        title: "\u521b\u5efa\u4f60\u7684\u7b2c\u4e00\u4e2a 智能体",
         description:
           "\u7ed9\u5b83\u8d77\u4e2a\u540d\u5b57\uff0c\u5199\u597d\u6307\u4ee4\uff0c\u9644\u52a0\u6280\u80fd\uff0c\u8bbe\u7f6e\u89e6\u53d1\u5668\u3002\u9009\u62e9\u5b83\u4f55\u65f6\u6fc0\u6d3b\uff1a\u88ab\u6307\u6d3e\u65f6\u3001\u6709\u8bc4\u8bba\u65f6\u3001\u88ab @\u63d0\u53ca\u65f6\u3002",
       },
       {
         title: "\u6307\u6d3e\u4e00\u4e2a Issue \u5e76\u89c2\u5bdf\u5b83\u5de5\u4f5c",
         description:
-          "\u4ece\u6307\u6d3e\u4eba\u4e0b\u62c9\u83dc\u5355\u4e2d\u9009\u62e9\u4f60\u7684 Agent\u2014\u2014\u5c31\u50cf\u6307\u6d3e\u7ed9\u540c\u4e8b\u4e00\u6837\u3002\u4efb\u52a1\u81ea\u52a8\u5165\u961f\u3001\u9886\u53d6\u3001\u6267\u884c\u3002\u5b9e\u65f6\u89c2\u770b\u8fdb\u5ea6\u3002",
+          "\u4ece\u6307\u6d3e\u4eba\u4e0b\u62c9\u83dc\u5355\u4e2d\u9009\u62e9\u4f60\u7684 智能体\u2014\u2014\u5c31\u50cf\u6307\u6d3e\u7ed9\u540c\u4e8b\u4e00\u6837\u3002\u4efb\u52a1\u81ea\u52a8\u5165\u961f\u3001\u9886\u53d6\u3001\u6267\u884c\u3002\u5b9e\u65f6\u89c2\u770b\u8fdb\u5ea6\u3002",
       },
     ],
     cta: "\u5f00\u59cb\u4f7f\u7528",
@@ -152,7 +152,7 @@ export function createZhDict(allowSignup: boolean): LandingDict {
     headlineLine1: "\u5f00\u6e90",
     headlineLine2: "\u4e3a\u6240\u6709\u4eba\u3002",
     description:
-      "Multica \u5b8c\u5168\u5f00\u6e90\u3002\u5ba1\u67e5\u6bcf\u4e00\u884c\u4ee3\u7801\uff0c\u6309\u4f60\u7684\u65b9\u5f0f\u81ea\u6258\u7ba1\uff0c\u5851\u9020\u4eba\u7c7b + Agent \u534f\u4f5c\u7684\u672a\u6765\u3002",
+      "Multica \u5b8c\u5168\u5f00\u6e90\u3002\u5ba1\u67e5\u6bcf\u4e00\u884c\u4ee3\u7801\uff0c\u6309\u4f60\u7684\u65b9\u5f0f\u81ea\u6258\u7ba1\uff0c\u5851\u9020\u4eba\u7c7b + 智能体 \u534f\u4f5c\u7684\u672a\u6765\u3002",
     cta: "\u5728 GitHub \u4e0a Star",
     highlights: [
       {
@@ -163,17 +163,17 @@ export function createZhDict(allowSignup: boolean): LandingDict {
       {
         title: "\u65e0\u4f9b\u5e94\u5546\u9501\u5b9a",
         description:
-          "\u81ea\u5e26 LLM \u63d0\u4f9b\u5546\u3001\u66f4\u6362 Agent \u540e\u7aef\u3001\u6269\u5c55 API\u3002\u4f60\u62e5\u6709\u6574\u4e2a\u6280\u672f\u6808\u7684\u63a7\u5236\u6743\u3002",
+          "\u81ea\u5e26 LLM \u63d0\u4f9b\u5546\u3001\u66f4\u6362 智能体 \u540e\u7aef\u3001\u6269\u5c55 API\u3002\u4f60\u62e5\u6709\u6574\u4e2a\u6280\u672f\u6808\u7684\u63a7\u5236\u6743\u3002",
       },
       {
         title: "\u9ed8\u8ba4\u900f\u660e",
         description:
-          "\u6bcf\u4e00\u884c\u4ee3\u7801\u90fd\u53ef\u5ba1\u8ba1\u3002\u786e\u5207\u4e86\u89e3\u4f60\u7684 Agent \u5982\u4f55\u505a\u51b3\u7b56\u3001\u4efb\u52a1\u5982\u4f55\u8def\u7531\u3001\u6570\u636e\u6d41\u5411\u4f55\u65b9\u3002",
+          "\u6bcf\u4e00\u884c\u4ee3\u7801\u90fd\u53ef\u5ba1\u8ba1\u3002\u786e\u5207\u4e86\u89e3\u4f60\u7684 智能体 \u5982\u4f55\u505a\u51b3\u7b56\u3001\u4efb\u52a1\u5982\u4f55\u8def\u7531\u3001\u6570\u636e\u6d41\u5411\u4f55\u65b9\u3002",
       },
       {
         title: "\u793e\u533a\u9a71\u52a8",
         description:
-          "\u4e0e\u793e\u533a\u4e00\u8d77\u5efa\u8bbe\uff0c\u800c\u4e0d\u4ec5\u4ec5\u662f\u4e3a\u793e\u533a\u5efa\u8bbe\u3002\u8d21\u732e\u6280\u80fd\u3001\u96c6\u6210\u548c Agent \u540e\u7aef\uff0c\u8ba9\u6bcf\u4e2a\u4eba\u53d7\u76ca\u3002",
+          "\u4e0e\u793e\u533a\u4e00\u8d77\u5efa\u8bbe\uff0c\u800c\u4e0d\u4ec5\u4ec5\u662f\u4e3a\u793e\u533a\u5efa\u8bbe\u3002\u8d21\u732e\u6280\u80fd\u3001\u96c6\u6210\u548c 智能体 \u540e\u7aef\uff0c\u8ba9\u6bcf\u4e2a\u4eba\u53d7\u76ca\u3002",
       },
     ],
   },
@@ -183,9 +183,9 @@ export function createZhDict(allowSignup: boolean): LandingDict {
     headline: "\u95ee\u4e0e\u7b54\u3002",
     items: [
       {
-        question: "Multica \u652f\u6301\u54ea\u4e9b\u7f16\u7801 Agent\uff1f",
+        question: "Multica \u652f\u6301\u54ea\u4e9b\u7f16\u7801 智能体\uff1f",
         answer:
-          "Multica \u76ee\u524d\u5f00\u7bb1\u5373\u7528\u652f\u6301 Claude Code\u3001Codex\u3001OpenClaw \u548c OpenCode\u3002\u5b88\u62a4\u8fdb\u7a0b\u81ea\u52a8\u68c0\u6d4b\u4f60\u5b89\u88c5\u7684 CLI\u3002\u56e0\u4e3a\u5f00\u6e90\uff0c\u4f60\u4e5f\u53ef\u4ee5\u81ea\u5df1\u6dfb\u52a0\u540e\u7aef\u3002",
+          "Multica \u5f00\u7bb1\u5373\u7528\u652f\u6301 11 \u6b3e AI \u7f16\u7a0b\u5de5\u5177\uff1aClaude Code\u3001Codex\u3001Cursor\u3001Copilot\u3001Gemini\u3001Hermes\u3001Kimi\u3001Kiro CLI\u3001OpenCode\u3001OpenClaw\u3001Pi\u3002\u5b88\u62a4\u8fdb\u7a0b\u4f1a\u81ea\u52a8\u68c0\u6d4b\u672c\u673a\u5df2\u5b89\u88c5\u7684 CLI \u5e76\u4e3a\u6bcf\u6b3e\u6ce8\u518c\u4e00\u4e2a\u8fd0\u884c\u65f6\u3002\u56e0\u4e3a\u5f00\u6e90\uff0c\u4f60\u4e5f\u53ef\u4ee5\u81ea\u5df1\u6dfb\u52a0\u540e\u7aef\u3002",
       },
       {
         question: "\u9700\u8981\u81ea\u6258\u7ba1\u5417\uff0c\u8fd8\u662f\u6709\u4e91\u7248\u672c\uff1f",
@@ -194,31 +194,31 @@ export function createZhDict(allowSignup: boolean): LandingDict {
       },
       {
         question:
-          "\u8fd9\u548c\u76f4\u63a5\u7528\u7f16\u7801 Agent \u6709\u4ec0\u4e48\u533a\u522b\uff1f",
+          "\u8fd9\u548c\u76f4\u63a5\u7528\u7f16\u7801 智能体 \u6709\u4ec0\u4e48\u533a\u522b\uff1f",
         answer:
-          "\u7f16\u7801 Agent \u64c5\u957f\u6267\u884c\u3002Multica \u6dfb\u52a0\u7684\u662f\u7ba1\u7406\u5c42\uff1a\u4efb\u52a1\u961f\u5217\u3001\u56e2\u961f\u534f\u4f5c\u3001\u6280\u80fd\u590d\u7528\u3001\u8fd0\u884c\u65f6\u76d1\u63a7\uff0c\u4ee5\u53ca\u6bcf\u4e2a Agent \u5728\u505a\u4ec0\u4e48\u7684\u7edf\u4e00\u89c6\u56fe\u3002\u628a\u5b83\u60f3\u8c61\u6210\u4f60\u7684 Agent \u7684\u9879\u76ee\u7ecf\u7406\u3002",
+          "\u7f16\u7801 智能体 \u64c5\u957f\u6267\u884c\u3002Multica \u6dfb\u52a0\u7684\u662f\u7ba1\u7406\u5c42\uff1a\u4efb\u52a1\u961f\u5217\u3001\u56e2\u961f\u534f\u4f5c\u3001\u6280\u80fd\u590d\u7528\u3001\u8fd0\u884c\u65f6\u76d1\u63a7\uff0c\u4ee5\u53ca\u6bcf\u4e2a 智能体 \u5728\u505a\u4ec0\u4e48\u7684\u7edf\u4e00\u89c6\u56fe\u3002\u628a\u5b83\u60f3\u8c61\u6210\u4f60\u7684 智能体 \u7684\u9879\u76ee\u7ecf\u7406\u3002",
       },
       {
-        question: "Agent \u80fd\u81ea\u4e3b\u5904\u7406\u957f\u65f6\u95f4\u4efb\u52a1\u5417\uff1f",
+        question: "智能体 \u80fd\u81ea\u4e3b\u5904\u7406\u957f\u65f6\u95f4\u4efb\u52a1\u5417\uff1f",
         answer:
-          "\u53ef\u4ee5\u3002Multica \u7ba1\u7406\u5b8c\u6574\u7684\u4efb\u52a1\u751f\u547d\u5468\u671f\u2014\u2014\u5165\u961f\u3001\u9886\u53d6\u3001\u6267\u884c\u3001\u5b8c\u6210\u6216\u5931\u8d25\u3002Agent \u4e3b\u52a8\u62a5\u544a\u963b\u585e\u5e76\u5b9e\u65f6\u63a8\u9001\u8fdb\u5ea6\u3002\u4f60\u53ef\u4ee5\u968f\u65f6\u67e5\u770b\uff0c\u4e5f\u53ef\u4ee5\u8ba9\u5b83\u4eec\u8fd0\u884c\u6574\u665a\u3002",
+          "\u53ef\u4ee5\u3002Multica \u7ba1\u7406\u5b8c\u6574\u7684\u4efb\u52a1\u751f\u547d\u5468\u671f\u2014\u2014\u5165\u961f\u3001\u9886\u53d6\u3001\u6267\u884c\u3001\u5b8c\u6210\u6216\u5931\u8d25\u3002智能体 \u4e3b\u52a8\u62a5\u544a\u963b\u585e\u5e76\u5b9e\u65f6\u63a8\u9001\u8fdb\u5ea6\u3002\u4f60\u53ef\u4ee5\u968f\u65f6\u67e5\u770b\uff0c\u4e5f\u53ef\u4ee5\u8ba9\u5b83\u4eec\u8fd0\u884c\u6574\u665a\u3002",
       },
       {
-        question: "\u6211\u7684\u4ee3\u7801\u5b89\u5168\u5417\uff1fAgent \u5728\u54ea\u91cc\u6267\u884c\uff1f",
+        question: "\u6211\u7684\u4ee3\u7801\u5b89\u5168\u5417\uff1f智能体 \u5728\u54ea\u91cc\u6267\u884c\uff1f",
         answer:
-          "Agent \u5728\u4f60\u7684\u673a\u5668\uff08\u672c\u5730\u5b88\u62a4\u8fdb\u7a0b\uff09\u6216\u4f60\u81ea\u5df1\u7684\u4e91\u57fa\u7840\u8bbe\u65bd\u4e0a\u6267\u884c\u3002\u4ee3\u7801\u6c38\u8fdc\u4e0d\u4f1a\u7ecf\u8fc7 Multica \u670d\u52a1\u5668\u3002\u5e73\u53f0\u53ea\u534f\u8c03\u4efb\u52a1\u72b6\u6001\u548c\u5e7f\u64ad\u4e8b\u4ef6\u3002",
+          "智能体 \u5728\u4f60\u7684\u673a\u5668\uff08\u672c\u5730\u5b88\u62a4\u8fdb\u7a0b\uff09\u6216\u4f60\u81ea\u5df1\u7684\u4e91\u57fa\u7840\u8bbe\u65bd\u4e0a\u6267\u884c\u3002\u4ee3\u7801\u6c38\u8fdc\u4e0d\u4f1a\u7ecf\u8fc7 Multica \u670d\u52a1\u5668\u3002\u5e73\u53f0\u53ea\u534f\u8c03\u4efb\u52a1\u72b6\u6001\u548c\u5e7f\u64ad\u4e8b\u4ef6\u3002",
       },
       {
-        question: "\u6211\u53ef\u4ee5\u8fd0\u884c\u591a\u5c11\u4e2a Agent\uff1f",
+        question: "\u6211\u53ef\u4ee5\u8fd0\u884c\u591a\u5c11\u4e2a 智能体\uff1f",
         answer:
-          "\u53d6\u51b3\u4e8e\u4f60\u7684\u786c\u4ef6\u3002\u6bcf\u4e2a Agent \u6709\u53ef\u914d\u7f6e\u7684\u5e76\u53d1\u9650\u5236\uff0c\u4f60\u53ef\u4ee5\u8fde\u63a5\u591a\u53f0\u673a\u5668\u4f5c\u4e3a\u8fd0\u884c\u65f6\u3002\u5f00\u6e90\u7248\u672c\u6ca1\u6709\u4efb\u4f55\u4eba\u4e3a\u9650\u5236\u3002",
+          "\u53d6\u51b3\u4e8e\u4f60\u7684\u786c\u4ef6\u3002\u6bcf\u4e2a 智能体 \u6709\u53ef\u914d\u7f6e\u7684\u5e76\u53d1\u9650\u5236\uff0c\u4f60\u53ef\u4ee5\u8fde\u63a5\u591a\u53f0\u673a\u5668\u4f5c\u4e3a\u8fd0\u884c\u65f6\u3002\u5f00\u6e90\u7248\u672c\u6ca1\u6709\u4efb\u4f55\u4eba\u4e3a\u9650\u5236\u3002",
       },
     ],
   },
 
   footer: {
     tagline:
-      "\u4eba\u7c7b + Agent \u56e2\u961f\u7684\u9879\u76ee\u7ba1\u7406\u3002\u5f00\u6e90\u3001\u53ef\u81ea\u6258\u7ba1\u3001\u4e3a\u672a\u6765\u7684\u5de5\u4f5c\u65b9\u5f0f\u800c\u5efa\u3002",
+      "\u4eba\u7c7b + 智能体 \u56e2\u961f\u7684\u9879\u76ee\u7ba1\u7406\u3002\u5f00\u6e90\u3001\u53ef\u81ea\u6258\u7ba1\u3001\u4e3a\u672a\u6765\u7684\u5de5\u4f5c\u65b9\u5f0f\u800c\u5efa\u3002",
     cta: "\u5f00\u59cb\u4f7f\u7528",
     groups: {
       product: {
@@ -283,6 +283,165 @@ export function createZhDict(allowSignup: boolean): LandingDict {
       fixes: "问题修复",
     },
     entries: [
+      {
+        version: "0.2.28",
+        date: "2026-05-08",
+        title: "Daemon 磁盘占用 CLI、Timeline 打磨与任务用量聚合提速",
+        changes: [],
+        features: [
+          "新增 `multica daemon disk-usage` CLI，按 task / workspace 维度查看磁盘占用",
+          "Skill Picker 弹窗新增搜索框，Agent 设置里挑技能更快",
+          "Daemon GC 覆盖扩展到 chat、autopilot、quick-create 任务",
+          "Issue 详情页面包屑直接显示 MUL-xxxx identifier",
+        ],
+        improvements: [
+          "Timeline 分页 size 提到 50，评论与活动按池独立 keyset 游标，长 Issue 翻页更顺",
+          "Show older / newer 按钮在边界场景也能正确出现，且视觉上更明显是可点击的",
+          "服务端 `task_usage` 聚合到每日 rollup 表，DB 负载明显下降",
+          "Daemon health check 在 repo 查询时不再阻塞，始终保持响应",
+          "Runtime 统计排除已归档的 agent，活跃数字更准",
+        ],
+        fixes: [
+          "Linux 上 daemon self-restart 改走 `brew prefix` 软链，Homebrew Cellar 删除后不再让 runtime 失联",
+          "CLI 短 ID 现在可以正确路由，复制粘贴的短前缀不再 404",
+          "Windows 上非 ASCII 字符评论 / 描述输入新增 `--content-file` / `--description-file`",
+          "Windows / Linux 桌面端用 Multica asterisk 替换 Electron 默认占位图标",
+          "Timeline 中孤立的 reply 现在会被正确捞回展示",
+          "Timeline 评论分页预算不再把 activity 算进去，避免活动多时挤掉真实评论",
+        ],
+      },
+      {
+        version: "0.2.27",
+        date: "2026-05-07",
+        title: "Chat 更顺手，Skill 支持 GitHub 导入，稳定性更好",
+        changes: [],
+        features: [
+          "支持直接通过 GitHub 链接导入可复用 Skill",
+        ],
+        improvements: [
+          "Chat 和 Inbox 更顺手，历史更清晰，复制回复更方便，归档后能更快处理下一项",
+          "Issue 操作会保留更多上下文，例如更容易找到对应本地文件夹，子 Issue 也会带上正确的项目和状态",
+          "Autopilot 连续失败后会自动暂停，异常自动化更容易发现和修复",
+        ],
+        fixes: [
+          "中文输入、桌面端升级、长 Issue 时间线和实时状态展示更稳定",
+        ],
+      },
+      {
+        version: "0.2.26",
+        date: "2026-05-06",
+        title: "i18n 全量铺开、长 Issue Timeline 提速与系统通知开关",
+        changes: [],
+        features: [
+          "Web 端完成简中翻译，21 个命名空间齐全，语言偏好按账号同步",
+          "Settings 新增 System Notifications 开关",
+          "支持删除 Chat 会话，History 面板移至 chat header",
+          "Runtime 在线判断改走 Redis（DB 兜底）",
+          "Desktop 支持加载 runtime 自托管配置",
+          "CLI 新增 `--assignee-id` / `--to-id` / `--user-id`，重名时定位更准",
+        ],
+        improvements: [
+          "Settings 的 Appearance Tab 改名为 Preferences，并把当前激活的 Tab 反映到 URL，深链可分享",
+          "长 Issue 打开秒开 —— Timeline 改为基于游标的 keyset 分页，重复的 `task_completed` / `task_failed` 活动条目合并展示",
+          "Runtime poll 与 heartbeat 调度按 runtime 隔离，单个忙碌 runtime 不再拖慢其他",
+          "CLI 更新请求落 Redis，server 重启也不丢",
+          "Runtime 用量统计窗口由 180 天收窄到 14 天，降低查询压力",
+          "项目列表返回 `resource_count` 摘要，不再内联全部 resource，响应体更小",
+          "404 页面重新设计，并修复 No-Access 重定向死循环",
+          "Quick Create 对 git-describe 类 daemon 跳过 CLI 版本闸",
+          "CI 启用 lint 强制门禁，历史 lint 债同步清理完毕",
+        ],
+        fixes: [
+          "Task 在服务端被删后，daemon 主动取消正在运行的 agent，避免孤儿进程",
+          "复用 execenv 时刷新陈旧的 Codex `auth.json`，修复偶发鉴权失败",
+          "`issue_id` 为空时拒绝写入 `.gc_meta.json`",
+          "跨 ACP 后端的 session/resume 信任 agent 自报的 session id，修复串号问题",
+          "OpenCode 的 skills 写到 `.opencode/skills/` 让其原生发现",
+          "Daemon 对 task-not-found 的 404 语义在 server 和最终 guard 双重收紧",
+          "侧边栏中失效的 Pin 自动取消挂载",
+          "项目详情页桌面端与移动端侧边栏状态独立保存",
+          "Runtime 详情页隐藏已归档的 agent",
+          "Add Resource 列表中已挂载的 repo 显示 URL tooltip；空项目页加上 New Issue 入口",
+          "S3 公开 URL 携带 region，修复跨区访问失败",
+          "Windows 安装器修正版本号解析与 checksum 解码",
+          "Quick Create 提交按钮去掉重复的快捷键提示",
+        ],
+      },
+      {
+        version: "0.2.24",
+        date: "2026-05-03",
+        title: "Repo Checkout `--ref`、Hermes 历史回放修复与多副本 Model Picker",
+        changes: [],
+        features: [
+          "`multica repo checkout --ref` 支持按分支、tag 或指定 commit 拉取仓库",
+          "`multica agent avatar` 命令支持直接通过 CLI 上传 Agent 头像",
+          "Inbox 中已完成任务新增 archive 按钮，移除冗余的 mark-as-done 悬浮按钮",
+        ],
+        improvements: [
+          "长 timeline 的 Issue 从 Inbox 打开不再卡顿 —— Markdown 渲染管线已 memoize，无关的 WS 事件不会再重渲染数千条评论",
+          "Model Picker 在多副本部署下可用 —— pending 请求改走 Redis 持久化，Daemon 上报失败也会自动重试",
+          "Daemon 空认领缓存 TTL 调高，空闲态 DB 压力进一步下降",
+        ],
+        fixes: [
+          "新创建的 Agent 立刻在各处可见 —— 创建时即 hydrate Agent 缓存",
+          "Hermes 在新一轮对话开始时不再重放上一轮答案 —— 历史 chunk 受单轮门禁限制",
+          "Codex runtime 模型选择器开放 GPT-5.5 系列",
+          "`multica login --token <PAT>` 正确接收 PAT 作为参数值",
+          "CLI update 完成状态上报更可靠",
+          "Session resume 按 runtime 正确守卫，避免跨 runtime 复用 session",
+          "看板拖拽 Issue 时显示设置不再丢失",
+          "Autopilot 列表在移动端 viewport 下响应式排版",
+          "Quick Create 生成的描述更贴合用户输入",
+          "Skill upsert 清理 null bytes，修复 PostgreSQL UTF8 错误",
+          "Connect Remote 弹窗的安装脚本 URL 修正",
+        ],
+      },
+      {
+        version: "0.2.21",
+        date: "2026-04-30",
+        title: "Quick Capture 全面升级、Mermaid 图表与 Typed Project Resources",
+        changes: [],
+        features: [
+          "Quick Capture 取代旧的 New Issue 弹窗 —— 支持连续创建、文件上传，并能根据粘贴的 URL 自动丰富标题与描述",
+          "Markdown 内联渲染 Mermaid 图表，复杂图支持全屏 lightbox",
+          "Project 支持单独绑定 repo，无需依赖 workspace 默认配置",
+          "Agent / 评论 / Runtime / Skill 全面接入权限感知 UI，没有权限的操作不再展示",
+        ],
+        improvements: [
+          "Daemon `/tasks/claim` 轮询走 Redis 空认领 fast-path，空闲态 DB 压力下降，长期 open 的 Issue 自动回收磁盘",
+          "Multica Agent 的 Git 提交自动追加 `Co-authored-by` trailer，归属更清晰",
+          "Desktop 拦截 Cmd+R / Ctrl+R / F5 防止意外刷新，开发模式与 Updates 设置中均展示真实版本号",
+        ],
+        fixes: [
+          "Quick Create 不再凭空脑补需求，并自动把发起人订阅到 Issue",
+          "Inbox 点击通知后立即跳到目标评论；从 Issue 详情页 Mark as Done 时自动归档",
+          "Task rerun 启动全新 session，跳过被污染的 resume 状态",
+          "受邀成员登录后路由到所在 workspace，不再强制带去 `/onboarding`",
+        ],
+      },
+      {
+        version: "0.2.20",
+        date: "2026-04-29",
+        title: "Create Issue by Agent、Agent Presence v3 与 Daemon WebSocket 心跳",
+        changes: [],
+        features: [
+          "Create Issue by Agent —— 按 `c` 输入一句话并选 Agent，Issue 异步创建，结果回执送达 Inbox",
+          "Agent Presence v3 —— 可用性与最近任务拆成两条更清晰的信号；Issue 详情右侧新增 Execution Log，可看到当前 active run 与历史 run",
+          "Daemon ↔ Server 心跳改走 WebSocket，HTTP 自动 fallback，任务起跑延迟更低",
+          "Mention 选择器按本机最近使用排序",
+        ],
+        improvements: [
+          "Server 用 Redis 缓存 PAT / Daemon Token 校验，大型团队不再让 DB 抗下每次请求",
+          "后端支持通过 `MULTICA_CLAUDE_ARGS` / `MULTICA_CODEX_ARGS` 配置 Agent CLI 默认参数",
+          "Manual 与 Agent 创建 Issue 共享同一个 Dialog 外壳，picker Agent 会被默认设为 assignee",
+        ],
+        fixes: [
+          "Create Issue by Agent 不再卡住 queued 任务，也不再因附件上传失败而重复创建 Issue",
+          "Agent 评论保留换行，不再渲染成字面量 `\\n`，多行回复的格式也被完整保留",
+          "Agent 自身发出的根评论不再继承父评论的 @mention，避免互相唤起的死循环",
+          "Windows 下 Cursor Agent 启动时保留多行 prompt",
+        ],
+      },
       {
         version: "0.2.19",
         date: "2026-04-28",

apps/web/next-env.d.ts

@@ -1,6 +1,6 @@
 /// <reference types="next" />
 /// <reference types="next/image-types/global" />
-import "./.next/types/routes.d.ts";
+import "./.next/dev/types/routes.d.ts";
 
 // NOTE: This file should not be edited
 // see https://nextjs.org/docs/app/api-reference/config/typescript for more information.

apps/web/platform/navigation.tsx

@@ -22,6 +22,8 @@ function NavigationProviderInner({
     back: router.back,
     pathname,
     searchParams: new URLSearchParams(searchParams.toString()),
+    getShareableUrl: (path: string) =>
+      typeof window === "undefined" ? path : window.location.origin + path,
   };
 
   return <NavigationProvider value={adapter}>{children}</NavigationProvider>;

apps/web/proxy.ts

@@ -1,4 +1,5 @@
 import { NextResponse, type NextRequest } from "next/server";
+import { matchLocale, LOCALE_COOKIE } from "@multica/core/i18n";
 
 // Old workspace-scoped route segments that existed before the URL refactor
 // (pre-#1131). Any URL with these as the FIRST segment is a legacy URL that
@@ -16,7 +17,34 @@ const LEGACY_ROUTE_SEGMENTS = new Set([
   "settings",
 ]);
 
-// Next.js 16 renamed `middleware` → `proxy`. The runtime API is identical.
+// Resolve the active locale per request. Cookie wins over Accept-Language;
+// matchLocale() falls back to DEFAULT_LOCALE when neither yields a match.
+function resolveLocale(req: NextRequest): string {
+  const cookieLocale = req.cookies.get(LOCALE_COOKIE)?.value;
+  const acceptLanguage = req.headers.get("accept-language") ?? "";
+  const candidates: string[] = [];
+  if (cookieLocale) candidates.push(cookieLocale);
+  for (const part of acceptLanguage.split(",")) {
+    const tag = part.split(";")[0]?.trim();
+    if (tag) candidates.push(tag);
+  }
+  return matchLocale(candidates);
+}
+
+// Forward the resolved locale to RSC layouts via the `x-multica-locale`
+// request header. layout.tsx reads it through `await headers()`. The
+// `request: { headers }` form is what makes the header land on the upstream
+// request — without it the value would only sit on the response.
+function nextWithLocale(req: NextRequest): NextResponse {
+  const headers = new Headers(req.headers);
+  headers.set("x-multica-locale", resolveLocale(req));
+  return NextResponse.next({ request: { headers } });
+}
+
+// Next.js 16 renamed `middleware` → `proxy`. API surface (NextRequest /
+// NextResponse / cookies / matcher) is identical; the only behavioral
+// change is the runtime — proxy is forced to nodejs and cannot opt into
+// edge.
 export function proxy(req: NextRequest) {
   const { pathname } = req.nextUrl;
   const hasSession = req.cookies.has("multica_logged_in");
@@ -48,34 +76,21 @@ export function proxy(req: NextRequest) {
   }
 
   // --- Root path: redirect logged-in users to their last workspace ---
-  if (pathname === "/") {
-    if (!hasSession) return NextResponse.next();
-
-    if (lastSlug) {
-      const url = req.nextUrl.clone();
-      url.pathname = `/${lastSlug}/issues`;
-      return NextResponse.redirect(url);
-    }
-
-    // No last_workspace_slug cookie → let landing page pick the first workspace
-    // client-side (features/landing/components/redirect-if-authenticated.tsx).
-    return NextResponse.next();
+  if (pathname === "/" && hasSession && lastSlug) {
+    const url = req.nextUrl.clone();
+    url.pathname = `/${lastSlug}/issues`;
+    return NextResponse.redirect(url);
   }
 
-  return NextResponse.next();
+  // --- Default: forward locale header to RSC, no redirect/rewrite ---
+  // Covers logged-out root path, /login, /:slug/*, and everything else.
+  return nextWithLocale(req);
 }
 
 export const config = {
-  matcher: [
-    "/",
-    "/issues/:path*",
-    "/projects/:path*",
-    "/agents/:path*",
-    "/inbox/:path*",
-    "/my-issues/:path*",
-    "/autopilots/:path*",
-    "/runtimes/:path*",
-    "/skills/:path*",
-    "/settings/:path*",
-  ],
+  // i18n header must land on every page request, so we use the standard
+  // negative-lookahead pattern from Next's i18n guide: skip API routes
+  // (Go backend), Next internals, and any path with a file extension
+  // (favicons, sw.js, public/* assets).
+  matcher: ["/((?!api|_next/static|_next/image|favicon.ico|.*\\.).*)"],
 };

apps/web/test/helpers.tsx

@@ -14,6 +14,7 @@ export const mockUser: User = {
   // Matches real server behavior for anyone who onboarded before this
   // field shipped — migration 054 backfills 'skipped_legacy'.
   starter_content_state: "skipped_legacy",
+  language: null,
   created_at: "2026-01-01T00:00:00Z",
   updated_at: "2026-01-01T00:00:00Z",
 };

docs/analytics.md

@@ -14,6 +14,7 @@ All analytics shipping is toggled by environment variables (see `.env.example`):
 |---|---|---|
 | `POSTHOG_API_KEY` | PostHog project API key. Empty = no events are shipped. | `""` |
 | `POSTHOG_HOST` | PostHog host (US or EU cloud, or self-hosted URL). | `https://us.i.posthog.com` |
+| `ANALYTICS_ENVIRONMENT` | Optional override for the standard `environment` event property. Normalized to `production`, `staging`, or `dev`; defaults from `APP_ENV`. | `APP_ENV` / `dev` |
 | `ANALYTICS_DISABLED` | Set to `true`/`1` to force the no-op client even when `POSTHOG_API_KEY` is set. | `""` |
 
 Local dev and self-hosted instances run with `POSTHOG_API_KEY=""`, so **no
@@ -82,6 +83,50 @@ handler → analytics.Client.Capture(Event)   ← non-blocking, returns immediat
   `$set_once` only for values that must never be overwritten (email,
   initial attribution, first-completion timestamp).
 
+## Taxonomy
+
+Every event is assigned to one dashboard category:
+
+| Category | Events |
+|---|---|
+| `core_loop` | `workspace_created`, `runtime_registered`, `runtime_ready`, `runtime_failed`, `runtime_offline`, `agent_created`, `issue_created`, `chat_message_sent`, `agent_task_queued`, `agent_task_dispatched`, `agent_task_started`, `agent_task_completed`, `agent_task_failed`, `agent_task_cancelled`, `autopilot_run_started`, `autopilot_run_completed`, `autopilot_run_failed` |
+| `onboarding_support` | `onboarding_started`, `onboarding_questionnaire_submitted`, `onboarding_completed`, `onboarding_runtime_path_selected`, `onboarding_runtime_detected`, `starter_content_decided` |
+| `acquisition` | `signup`, `download_intent_expressed`, `download_page_viewed`, `download_initiated`, `cloud_waitlist_joined` |
+| `ops_feedback` | `feedback_opened`, `feedback_submitted` |
+| `system/noise` | `$pageview`, `$set`, `$identify`, `$autocapture`, `$rageclick` |
+
+The v0 core dashboard must use only `core_loop` plus the specific
+`onboarding_support` steps used by the activation funnel. Acquisition,
+feedback, and system/noise events stay in separate dashboards.
+
+## Standard core properties
+
+Canonical core events should carry these properties whenever the entity exists:
+
+| Property | Type | Notes |
+|---|---|---|
+| `environment` | string | `production` / `staging` / `dev`; stamped by backend and frontend analytics clients. |
+| `event_schema_version` | int | Current version: `2`. |
+| `user_id` | string UUID | Human user ID when known. Agent/system events may omit it. |
+| `workspace_id` | string UUID | Required for workspace-scoped events. |
+| `agent_id` | string UUID | Required for agent/task events. |
+| `task_id` | string UUID | Required for `agent_task_*` events. |
+| `issue_id` / `chat_session_id` / `autopilot_run_id` | string UUID | Relevant source entity for the task/entry event. |
+| `source` | string | Canonical values: `onboarding`, `manual`, `chat`, `autopilot`, `api`. UI surface details use `surface` or `trigger_source`. |
+| `runtime_mode` | string | `cloud` / `local` when a runtime/agent task is involved. |
+| `provider` | string | `claude`, `codex`, `cursor`, etc. when a runtime/agent task is involved. |
+| `is_demo` | bool | Currently always `false`; reserved for future demo/test workspace filtering. |
+
+Task terminal events additionally carry `duration_ms`; failures carry
+`failure_reason`, `error_type`, and `will_retry`. Runtime failure events carry
+`recoverable`; runtime ready events carry `runtime_id`, `ready_duration_ms`
+only when it is actually measured, and `daemon_id` for local runtimes.
+
+Schema v2 is the first canonical core-metrics schema. It replaces early v1
+drafts that mirrored `failure_reason` into `error_type`, used `recoverable`
+for task/autopilot failures, and emitted `ready_duration_ms: 0` before the
+registration path had a measured duration.
+
 ## Event contract
 
 ### `signup`
@@ -128,6 +173,8 @@ extra query, no race.
 | Property | Type | Description |
 |---|---|---|
 | `runtime_id` | string (UUID) | The newly created agent_runtime row id. |
+| `daemon_id` | string | Local daemon identity when available. |
+| `runtime_mode` | string | Currently `local`; reserved for cloud runtimes. |
 | `provider` | string | e.g. `"codex"`, `"claude"`. |
 | `runtime_version` | string | Version of the agent runtime binary. |
 | `cli_version` | string | Version of the `multica` CLI that registered it. |
@@ -137,6 +184,118 @@ registered via a member's JWT/PAT; daemon-token registrations fall back to
 `workspace:<workspace_id>` so PostHog doesn't bucket unrelated daemons
 under a single "anonymous" person.
 
+### `runtime_ready`
+
+Fires when a runtime is first registered in an online/ready state. This is the
+activation-funnel step that should replace treating `runtime_registered` as
+proof of readiness. The backend emits this only on the INSERT path for a new
+`agent_runtime` row; ordinary daemon reconnects update the existing row and do
+not emit another `runtime_ready`. Dashboard funnels should still count
+distinct `runtime_id`.
+
+| Property | Type | Description |
+|---|---|---|
+| `runtime_id` | string (UUID) | The `agent_runtime` row id. |
+| `daemon_id` | string | Local daemon identity when available. |
+| `ready_duration_ms` | int64 | Optional. Time from registration start to ready; omitted until the registration path can measure it. |
+| `runtime_mode` | string | `local` / `cloud`. |
+| `provider` | string | Runtime provider. |
+
+### `runtime_failed`
+
+Fires when runtime setup/registration fails before a ready runtime can be
+recorded. Today this is scoped to backend registration persistence failures;
+future setup flows should reuse it for provider detection or daemon boot
+failures.
+
+| Property | Type | Description |
+|---|---|---|
+| `daemon_id` | string | Local daemon identity when available. |
+| `provider` | string | Runtime provider attempted. |
+| `failure_reason` | string | Stable coarse reason. |
+| `error_type` | string | Stable error classifier. |
+| `recoverable` | bool | Whether retrying setup may succeed. |
+
+### `runtime_offline`
+
+Fires when a runtime is explicitly deregistered or the backend sweeper marks it
+offline after missed heartbeats. This is not an activation step; it supports
+local runtime retention and drop-off diagnosis.
+
+### `issue_created`
+
+Fires after an issue row is created, including manual UI/API issue creation,
+quick-create issue creation by an agent, and autopilot `create_issue` runs.
+
+| Property | Type | Description |
+|---|---|---|
+| `issue_id` | string (UUID) | Created issue. |
+| `agent_id` | string (UUID) | Agent assignee or creating agent when applicable. |
+| `task_id` | string (UUID) | Present for quick-create issue creation. |
+| `autopilot_run_id` | string (UUID) | Present for autopilot-created issues. |
+| `source` | string | `manual`, `api`, or `autopilot`. |
+
+### `chat_message_sent`
+
+Fires after a user chat message is persisted and the corresponding agent task
+is queued.
+
+| Property | Type | Description |
+|---|---|---|
+| `chat_session_id` | string (UUID) | Chat session. |
+| `task_id` | string (UUID) | Queued agent task. |
+| `agent_id` | string (UUID) | Chat agent. |
+| `source` | string | Always `chat`. |
+
+### `agent_task_queued` / `agent_task_dispatched` / `agent_task_started` / `agent_task_completed`
+
+Canonical task lifecycle events emitted from `agent_task_queue` state
+transitions. `agent_task_dispatched` fires when the backend claims a queued
+task for a runtime, before the daemon marks it running with
+`agent_task_started`. These events replace `issue_executed` for core loop
+success metrics and allow the activation funnel to split queue backlog from
+claim/start handoff.
+
+| Property | Type | Description |
+|---|---|---|
+| `task_id` | string (UUID) | `agent_task_queue.id`; required. |
+| `agent_id` | string (UUID) | Owning agent. |
+| `issue_id` | string (UUID) | Present for issue-linked tasks. |
+| `chat_session_id` | string (UUID) | Present for chat tasks. |
+| `autopilot_run_id` | string (UUID) | Present for run-only autopilot tasks. |
+| `source` | string | `manual`, `chat`, or `autopilot`. |
+| `runtime_mode` | string | `local` / `cloud`. |
+| `provider` | string | Runtime provider. |
+| `duration_ms` | int64 | Terminal events only; measured from `started_at` when available. |
+
+### `agent_task_failed` / `agent_task_cancelled`
+
+Terminal task lifecycle events. They use the same join fields as
+`agent_task_completed`. `agent_task_failed` also carries:
+
+| Property | Type | Description |
+|---|---|---|
+| `failure_reason` | string | Stable reason from `agent_task_queue.failure_reason`, default `agent_error`. |
+| `error_type` | string | Stable coarse classifier, e.g. `runtime`, `timeout`, `agent_output`, `cancelled`, `agent_error`. |
+| `will_retry` | bool | Whether the backend auto-retry policy will create another task attempt. |
+
+### `autopilot_run_started` / `autopilot_run_completed` / `autopilot_run_failed`
+
+Fires from `autopilot_run` lifecycle changes. `source` is always
+`autopilot`; the trigger origin is carried in `trigger_source` (`manual`,
+`schedule`, `webhook`, or `api`).
+
+| Property | Type | Description |
+|---|---|---|
+| `autopilot_id` | string (UUID) | Autopilot definition. |
+| `autopilot_run_id` | string (UUID) | Run row. |
+| `agent_id` | string (UUID) | Assigned agent. |
+| `trigger_source` | string | `manual`, `schedule`, `webhook`, or `api`. |
+| `duration_ms` | int64 | Terminal events only. |
+| `failure_reason` | string | Failed events only. |
+| `error_type` | string | Failed events only; stable coarse classifier such as `configuration`, `issue_terminal`, `dispatch_error`, `task_error`, or `autopilot_error`. |
+| `will_retry` | bool | Failed events only; currently `false` because autopilot retry cadence is owned by triggers/schedules. |
+
 ### `issue_executed`
 
 Fires **at most once per issue** — when the first task on that issue
@@ -149,6 +308,11 @@ distinct issues, not tasks.
 | Property | Type | Description |
 |---|---|---|
 | `issue_id` | string (UUID) | |
+| `task_id` | string (UUID) | Completing task. |
+| `agent_id` | string (UUID) | Completing agent. |
+| `source` | string | `manual`, `chat`, or `autopilot`. |
+| `runtime_mode` | string | `local` / `cloud`. |
+| `provider` | string | Runtime provider. |
 | `task_duration_ms` | int64 | Wall-clock time between `task.started_at` and `task.completed_at`. Zero when the task was created in a completed state (rare). |
 
 `distinct_id` prefers the issue's human creator so agent-executed events
@@ -165,6 +329,10 @@ emit `n=1`. PostHog answers the same question at query time via
 and funnel steps of the form "workspace has had ≥2 `issue_executed`
 events" are expressible without the property. No information is lost.
 
+Compatibility: `issue_executed` remains a historical compatibility event for
+old dashboards. New core-loop success dashboards should use
+`agent_task_completed` and filter by `source`/`issue_id` as needed.
+
 ### `team_invite_sent`
 
 Fires from `CreateInvitation` after the DB row is written.
@@ -188,6 +356,17 @@ accepted and the member row is inserted in the same transaction.
 `distinct_id` is the invitee's user id — this is the event that closes the
 expansion funnel.
 
+### `onboarding_started`
+
+Fires once when the onboarding shell mounts and the initial workspace list has
+resolved. Existing-workspace users carry `workspace_id`; brand-new users do
+not have a workspace yet.
+
+| Property | Type | Description |
+|---|---|---|
+| `workspace_id` | string (UUID) | Present only when the user already has a workspace. |
+| `source` | string | Always `onboarding`. |
+
 ### `onboarding_questionnaire_submitted`
 
 Fires on the first PatchOnboarding that transitions the user's
@@ -226,6 +405,7 @@ isolates the Step 4 signal from later agent additions.
 |---|---|---|
 | `agent_id` | string (UUID) | |
 | `provider` | string | Runtime provider the agent is bound to (`claude`, `codex`, etc). |
+| `runtime_mode` | string | Runtime mode copied from the bound runtime. |
 | `template` | string | Template slug used to seed the agent (`coding` / `planning` / `writing` / `assistant`). Empty when the caller didn't come from a template picker. |
 | `is_first_agent_in_workspace` | bool | `true` when the workspace had zero agents before this insert. |
 
@@ -241,7 +421,8 @@ which exit the user took.
 
 | Property | Type | Description |
 |---|---|---|
-| `completion_path` | string | One of `full` / `runtime_skipped` / `cloud_waitlist` / `skip_existing` / `unknown`. See below. |
+| `workspace_id` | string (UUID) | Present for workspace-linked onboarding completions. |
+| `completion_path` | string | One of `full` / `runtime_skipped` / `cloud_waitlist` / `skip_existing` / `invite_accept` / `unknown`. See below. |
 | `joined_cloud_waitlist` | bool | Derived from `user.cloud_waitlist_email`. Orthogonal to `completion_path` — a user may submit the waitlist form and still pick CLI. |
 
 Person properties set with `$set_once`:
@@ -256,6 +437,7 @@ Person properties set with `$set_once`:
 - `runtime_skipped` — Completed without connecting a runtime (user hit Skip in Step 3).
 - `cloud_waitlist` — Submitted the cloud waitlist form and skipped Step 3.
 - `skip_existing` — "I've done this before" from Welcome. The user already had a workspace.
+- `invite_accept` — Accepted at least one workspace invitation.
 - `unknown` — Legacy fallback when the client didn't send a path. Should stay near zero after rollout.
 
 ### `cloud_waitlist_joined`
@@ -314,11 +496,11 @@ request payload.
   `packages/views/onboarding/steps/step-platform-fork.tsx` when the web
   user clicks one of the three Step 3 fork cards (before any server
   call happens, so it's frontend-only). Properties: `path`
-  (`download_desktop` / `cli` / `cloud_waitlist`), `source` (`step3`;
-  literal today but reserved for future surfaces reusing this event),
-  `is_mac`. Also writes `platform_preference` (`web` / `desktop`) to
-  person properties so every subsequent event on the user can be
-  broken down by chosen platform. **Note**: semantic "download
+  (`download_desktop` / `cli` / `cloud_waitlist`), `source`
+  (`onboarding`), `surface` (`step3`), `workspace_id`, and `is_mac`.
+  Also writes `platform_preference` (`web` / `desktop`) to person
+  properties so every subsequent event on the user can be broken down
+  by chosen platform. **Note**: semantic "download
   intent" is now better served by `download_intent_expressed` below —
   `path: "download_desktop"` signals Step 3 path choice specifically,
   not actual download start.
@@ -334,8 +516,9 @@ request payload.
   `runtime_registered` is silent on that cohort. Splits
   `completion_path=runtime_skipped` into "had CLIs, skipped anyway"
   vs "no CLIs available, had no choice". Properties:
-  - `source`: `step3_desktop` (literal; reserved for a future web
-    emission under a different value).
+  - `source`: `onboarding`.
+  - `surface`: `step3_desktop`.
+  - `workspace_id`: current onboarding workspace.
   - `outcome`: `found` (at least one runtime registered before the
     5 s grace window expired) or `empty` (none registered by then).
   - `runtime_count`: number of runtimes visible to this user at
@@ -419,6 +602,38 @@ request payload.
   `JSON.stringify`, and the entire payload is dropped if it still exceeds
   512 chars. That way PostHog sees either intact JSON or nothing at all.
 
+## Reconciliation
+
+`agent_task_completed` is the canonical PostHog-side task success event. It
+should reconcile daily against the operational source of truth:
+
+```sql
+SELECT date_trunc('day', completed_at AT TIME ZONE 'UTC') AS day,
+       count(*) AS db_completed_tasks
+FROM agent_task_queue
+WHERE status = 'completed'
+  AND completed_at >= now() - interval '30 days'
+GROUP BY 1
+ORDER BY 1;
+```
+
+Equivalent HogQL:
+
+```sql
+SELECT toStartOfDay(timestamp) AS day,
+       count() AS posthog_completed_tasks
+FROM events
+WHERE event = 'agent_task_completed'
+  AND properties.environment = 'production'
+  AND timestamp >= now() - interval 30 day
+GROUP BY day
+ORDER BY day
+```
+
+The expected difference should be near zero. Allow a small delay window for
+PostHog ingestion and backend analytics queue drops; sustained drift means
+either an emission site is missing or PostHog shipping is unhealthy.
+
 ## Governance
 
 Before adding, renaming, or removing any event:

docs/product-overview.md

@@ -372,7 +372,7 @@ skill
 3. **注入**：当 agent 认领任务时，daemon 把挂载的 skill 内容写到任务工作目录的 **provider 原生位置**：
    - Claude Code → `.claude/skills/{name}/SKILL.md`
    - Codex → `CODEX_HOME/skills/{name}/`
-   - OpenCode → `.config/opencode/skills/{name}/SKILL.md`
+   - OpenCode → `.opencode/skills/{name}/SKILL.md`
    - Pi → `.pi/skills/{name}/SKILL.md`
    - Cursor → `.cursor/skills/{name}/SKILL.md`
    - GitHub Copilot → `.github/skills/{name}/SKILL.md`

docs/rfcs/0001-mention-dedup-policy.md

@@ -0,0 +1,130 @@
+# RFC: Per-mention agent task enqueue (drop @mention coalescing dedup)
+
+- Issue: [MUL-1913](mention://issue/9f54962b-e055-43eb-a649-1b16db52fea2)
+- Status: Accepted
+- Date: 2026-05-09
+
+## Background
+
+When a member @mentions an agent on an issue (or a member comments on an
+issue assigned to an agent), the trigger path enqueues an `agent_task_queue`
+row. Today both paths short-circuit when the same agent already has a
+`queued` or `dispatched` task on the same issue:
+
+- `server/internal/handler/comment.go` `enqueueMentionedAgentTasks` — @mention
+  trigger
+- `server/internal/handler/issue.go` `shouldEnqueueOnComment` — assignee-on-
+  comment trigger
+
+Both call `Queries.HasPendingTaskForIssueAndAgent` and skip enqueue when it
+returns true. The intent was a coalescing queue: rapid-fire comments fold
+into a single pending task, and when that task picks up it reads all the
+latest comments anyway.
+
+## Problem
+
+The coalescing model has three user-visible costs:
+
+1. **No UI feedback for the merged comment.** A second @mention does not
+   create a task, so no queued banner appears and there is no toast saying
+   "merged into pending task". Users perceive the @mention as lost.
+2. **Trigger comment provenance is lost.** Only the first trigger comment
+   is recorded on the task; subsequent triggers are not referenced by any
+   task. Auditing "what made this run happen" fails.
+3. **Distinct intents collapse.** When two @mentions live in different
+   threads with different requests ("add a test" vs. "fix copy"), folding
+   them into one task forces the agent to disambiguate, and the user cannot
+   cancel one without cancelling both.
+
+Different threads and different mention text are strong signals that the
+two triggers are distinct intents — coalescing throws that signal away.
+
+## Decision
+
+**Adopt option C: every @mention or assignee-comment trigger creates its
+own task.** No `(issue, agent)` dedup at enqueue time.
+
+Per-(issue, agent) execution stays serial because `ClaimAgentTask`
+(`server/pkg/db/queries/agent.sql`) refuses to dispatch a queued row when
+the same agent has another `dispatched` or `running` row on the same
+issue. Multiple queued rows pile up safely and drain in
+`(priority DESC, created_at ASC)` order. This is a coordination-side
+property — `FOR UPDATE SKIP LOCKED` locks the row being claimed, not the
+`(issue, agent)` key — and relies on the daemon today never invoking
+`ClaimAgentTask` concurrently for the same agent. Tightening that into a
+real DB-level guarantee (e.g. an advisory lock keyed on `(issue, agent)`)
+is out of scope for this RFC.
+
+### Mutual exclusion between on_comment and @mention paths
+
+Without `(issue, agent)` dedup at enqueue time, a single member comment
+that @mentions the assignee would otherwise enqueue twice with identical
+`trigger_comment_id`: once via the on_comment path
+(`shouldEnqueueOnComment` → `EnqueueTaskForIssue`) and once via the
+@mention path (`enqueueMentionedAgentTasks` → `EnqueueTaskForMention`).
+Same trick applies to a plain reply that inherits the assignee mention
+from the thread root.
+
+The on_comment gate gains a `commentMentionsAssignee` clause that uses the
+same effective-mention computation as the @mention path
+(`shouldInheritParentMentions` for inheritance). When the @mention path
+will enqueue for the assignee, on_comment skips. The two paths become
+mutually exclusive on a `(comment, assignee)` pair.
+
+### Considered alternatives
+
+- **A. Keep coalescing, add a UI hint** ("merged into pending task"). Fixes
+  visibility but not provenance and not the distinct-intent case.
+- **B. Allow up to N queued tasks per (issue, agent), coalesce above N.**
+  Combines the worst of both — still loses the Nth+1 trigger comment, and
+  introduces a magic number.
+- **C. No dedup, every trigger creates a task. (Chosen.)**
+
+## Out of scope
+
+- **True rapid-fire duplicate suppression.** A user double-clicks @ within
+  a second; both create tasks and the agent runs twice on identical
+  context. Acceptable cost — the agent reads its own previous comment in
+  the second run and can early-exit. We may revisit by having the
+  scheduler skip a queued task when "no relevant comments since the
+  previous task for this (issue, agent) completed", but that is a
+  follow-up, not a blocker for this RFC.
+- **Cross-agent dedup.** Different agents on the same issue continue to
+  run in parallel; nothing changes there.
+- **Queued banner UI.** Already shipped in
+  [MUL-1897](mention://issue/14fdefb4-3a36-4406-a840-1f6700ac95b5).
+
+## Implementation
+
+1. Remove the `HasPendingTaskForIssueAndAgent` short-circuit in
+   `enqueueMentionedAgentTasks`.
+2. Remove the `HasPendingTaskForIssueAndAgent` short-circuit in
+   `shouldEnqueueOnComment`. The function reduces to the
+   assignee-readiness check (`isAgentAssigneeReady` + non-backlog status).
+3. Add the `commentMentionsAssignee` clause to the on_comment gate so
+   the on_comment and @mention paths are mutually exclusive on a
+   `(comment, assignee)` pair (see "Mutual exclusion" above).
+4. Drop `HasPendingTaskForIssueAndAgent` and the unused
+   `HasPendingTaskForIssue` from `server/pkg/db/queries/agent.sql`. Re-run
+   `make sqlc`.
+5. Tests:
+   - `TestRepeatedMentionsEnqueueSeparateTasks` — two @mentions on an
+     unassigned issue produce two `queued` rows with distinct
+     `trigger_comment_id` values.
+   - `TestAssigneeMentionDoesNotDoubleEnqueue` — a member comment that
+     @mentions the assignee on an assigned issue produces exactly one
+     `queued` row (the mention path), not two.
+6. No migration needed. No frontend changes needed: the queued banner
+   already aggregates over `ListActiveTasksByIssue`, so multiple queued
+   rows render correctly.
+
+## Risks
+
+- **Cost:** users who comment frequently on agent-assigned issues will
+  trigger more runs than today. Mitigated by per-(issue, agent) serial
+  execution — no extra concurrency, just more sequential work — and by
+  the future "skip if no relevant comments" optimization noted above.
+- **Replay:** the second queued task reads issue state that the first
+  task may have already addressed. Agents already need to read recent
+  comments and judge whether work is still required; this RFC does not
+  change that contract.

package.json

@@ -13,7 +13,8 @@
     "test": "turbo test",
     "lint": "turbo lint",
     "clean": "turbo clean && rm -rf node_modules",
-    "ui:add": "cd packages/ui && npx shadcn@latest add"
+    "ui:add": "cd packages/ui && npx shadcn@latest add",
+    "generate:reserved-slugs": "node scripts/generate-reserved-slugs.mjs"
   },
   "packageManager": "pnpm@10.28.2",
   "pnpm": {

packages/core/agents/index.ts

@@ -5,3 +5,5 @@ export * from "./use-agent-presence";
 export * from "./use-agent-activity";
 export * from "./use-workspace-presence-prefetch";
 export * from "./constants";
+export * from "./visibility-label";
+export * from "./use-workspace-agent-availability";

packages/core/agents/use-workspace-agent-availability.ts

@@ -0,0 +1,60 @@
+"use client";
+
+import { useQuery } from "@tanstack/react-query";
+import { useWorkspaceId } from "../hooks";
+import { useAuthStore } from "../auth";
+import { agentListOptions, memberListOptions } from "../workspace/queries";
+import { canAssignAgentToIssue } from "../permissions";
+
+/**
+ * Three-state availability for "does the current user have any agent
+ * they can chat with in this workspace?".
+ *
+ * Why three states (not a boolean): the answer to "is there an agent?"
+ * lives on the server. Until the agent-list query resolves, the answer
+ * is genuinely *unknown*. Callers must distinguish "loading" from
+ * "confirmed empty" — collapsing them to a boolean causes UIs to flash
+ * disabled/empty states for the first few hundred ms after mount, even
+ * when the workspace actually has agents.
+ *
+ *   "loading"   — agent or member list still in flight (be neutral in UI)
+ *   "none"      — both queries resolved, user has zero assignable agents
+ *   "available" — at least one agent passes archive + visibility filters
+ */
+export type WorkspaceAgentAvailability = "loading" | "none" | "available";
+
+/**
+ * Mirrors the per-agent visibility/archived filter used by AssigneePicker
+ * and the chat agent dropdown, so the three pickers can never disagree on
+ * "is this agent reachable?".
+ *
+ * Members are queried because `canAssignAgentToIssue` reads the caller's
+ * role to decide visibility for `private` agents — without member data,
+ * a freshly-loaded agent list could still produce wrong answers.
+ */
+export function useWorkspaceAgentAvailability(): WorkspaceAgentAvailability {
+  const wsId = useWorkspaceId();
+  const userId = useAuthStore((s) => s.user?.id);
+  const { data: agents, isFetched: agentsFetched } = useQuery(
+    agentListOptions(wsId),
+  );
+  const { data: members, isFetched: membersFetched } = useQuery(
+    memberListOptions(wsId),
+  );
+
+  if (!agentsFetched || !membersFetched) return "loading";
+
+  const rawRole = members?.find((m) => m.user_id === userId)?.role;
+  const role =
+    rawRole === "owner" || rawRole === "admin" || rawRole === "member"
+      ? rawRole
+      : null;
+
+  const hasVisibleAgent = (agents ?? []).some(
+    (a) =>
+      !a.archived_at &&
+      canAssignAgentToIssue(a, { userId: userId ?? null, role }).allowed,
+  );
+
+  return hasVisibleAgent ? "available" : "none";
+}

packages/core/agents/visibility-label.ts

@@ -0,0 +1,31 @@
+import type { AgentVisibility } from "../types";
+
+/**
+ * Display labels for agent visibility. The DB stores `private` as the value
+ * but the UI surface name is "Personal" — better matches what the field
+ * actually means now that workspace admins can also assign private agents.
+ */
+export const VISIBILITY_LABEL: Record<AgentVisibility, string> = {
+  workspace: "Workspace",
+  private: "Personal",
+};
+
+/**
+ * Honest descriptions for assignability. The previous "Only you can assign"
+ * text was a lie — workspace owners and admins can assign private agents too
+ * (server `issue.go:1471-1490`).
+ */
+export const VISIBILITY_DESCRIPTION: Record<AgentVisibility, string> = {
+  workspace: "All members can assign",
+  private: "Only you and workspace admins can assign",
+};
+
+/** Tooltip suitable for read-only badges on hover/list rows. */
+export const VISIBILITY_TOOLTIP: Record<AgentVisibility, string> = {
+  workspace: "Workspace — all members can assign",
+  private: "Personal — only you and workspace admins can assign",
+};
+
+export function visibilityLabel(v: AgentVisibility): string {
+  return VISIBILITY_LABEL[v];
+}

packages/core/analytics/index.test.ts

@@ -45,20 +45,33 @@ describe("initAnalytics super-properties", () => {
     expect(posthog.register).toHaveBeenCalledWith({
       client_type: "web",
       app_version: "1.2.3",
+      environment: "dev",
+      event_schema_version: 2,
+      is_demo: false,
     });
   });
 
   it("omits app_version when not provided", async () => {
     const { analytics, posthog } = await loadModule();
     analytics.initAnalytics({ key: "k", host: "" });
-    expect(posthog.register).toHaveBeenCalledWith({ client_type: "web" });
+    expect(posthog.register).toHaveBeenCalledWith({
+      client_type: "web",
+      environment: "dev",
+      event_schema_version: 2,
+      is_demo: false,
+    });
   });
 
   it("detects desktop when window.electron is present", async () => {
     vi.stubGlobal("window", { electron: {} });
     const { analytics, posthog } = await loadModule();
     analytics.initAnalytics({ key: "k", host: "" });
-    expect(posthog.register).toHaveBeenCalledWith({ client_type: "desktop" });
+    expect(posthog.register).toHaveBeenCalledWith({
+      client_type: "desktop",
+      environment: "dev",
+      event_schema_version: 2,
+      is_demo: false,
+    });
   });
 });
 
@@ -76,6 +89,9 @@ describe("resetAnalytics", () => {
     expect(posthog.register).toHaveBeenCalledWith({
       client_type: "web",
       app_version: "1.2.3",
+      environment: "dev",
+      event_schema_version: 2,
+      is_demo: false,
     });
   });