mirror of
https://github.com/multica-ai/multica.git
synced 2026-08-12 10:59:06 +02:00
c98161b039
Adds a dedicated bilingual /docs/squads page covering the squad model (leader + members), assignment, comment trigger rules, archive semantics, and the squad CLI surface. Wires the new page into meta.json and meta.zh.json under the Agents section, and adds short cross-references from agents, assigning-issues, mentioning-agents, and the CLI reference so users can discover squads from the pages they're already on. MUL-2206 Co-authored-by: multica-agent <github@multica.ai>
64 lines
4.3 KiB
Plaintext
64 lines
4.3 KiB
Plaintext
---
|
|
title: "@-mention agents in comments"
|
|
description: Mention an agent with @ to have it take a look from a comment — no assignee change, no status change, lighter than assigning.
|
|
---
|
|
|
|
import { Callout } from "fumadocs-ui/components/callout";
|
|
|
|
`@`-mentioning an [agent](/agents) in a [comment](/comments) is the lighter trigger — **no assignee change, no status change**, just a nudge to have the agent take a look at the current [issue](/issues). Compared to [**assigning**](/assigning-issues) (turning the agent into the owner and handing over the issue), @-mention fits "take a look at this section," "give me another angle," or "pull them in for a quick discussion."
|
|
|
|
## Mention an agent in a comment
|
|
|
|
Same as mentioning a member — type `@` to open the picker and select an agent. Once the comment is posted, Multica immediately enqueues a `task` for each mentioned agent with **that comment** as its trigger context. When the agent receives the task it can read:
|
|
|
|
- The full issue (description + every historical comment)
|
|
- The trigger comment itself — as the starting point for this run
|
|
|
|
The `@mention` Markdown syntax, the picker, and `@all` semantics are covered in [**Comments**](/comments).
|
|
|
|
<Callout type="info">
|
|
**You can also `@`-mention a [squad](/squads) in a comment.** The same picker surfaces squads alongside members and agents; selecting one inserts `[@SquadName](mention://squad/<uuid>)` and triggers the squad's **leader agent** to coordinate a response — assignee and status stay untouched.
|
|
</Callout>
|
|
|
|
## How it differs from assignment
|
|
|
|
Both put the agent to work, but the mechanics are entirely different:
|
|
|
|
| Dimension | Assign | @-mention |
|
|
|---|---|---|
|
|
| Changes `assignee` | ✓ | ✗ |
|
|
| Changes `status` | ✗ | ✗ |
|
|
| Enqueues a `task` | Immediately (non-Backlog) | Immediately |
|
|
| Trigger comment ID | Optional | Always carries the current comment |
|
|
| Agents targeted per action | 1 (one assignee) | Many (a comment can @ multiple) |
|
|
| Priority | Inherits from issue | Inherits from issue |
|
|
|
|
The rule of thumb is simple: **use assignment when you want the agent to "own this issue from now on"; use @-mention when you want it to "take a look at the current context."**
|
|
|
|
## What happens when you @ multiple agents
|
|
|
|
If one comment @-mentions several agents, each one is enqueued an independent `task` on its own runtime — **they run in parallel** without blocking each other.
|
|
|
|
If an agent already has a `queued` or `dispatched` `task` on the same issue (for example, it was just mentioned and has not started yet), the new mention is **deduplicated** and no duplicate `task` is enqueued. Deduplication is **scoped to a single comment** — two different comments seconds apart that both @ the same agent will both enqueue a `task`.
|
|
|
|
<Callout type="warning">
|
|
**Adding an @ by editing a comment does not re-trigger.** If you remember to add `@agent` only after posting, editing in the `@` only changes what is displayed — it **does not** deliver a new `task` to that agent. To trigger it, post a new comment or assign the issue to it.
|
|
</Callout>
|
|
|
|
## `@all` does not trigger any agent
|
|
|
|
When you call everyone with `@all`, **only workspace members land in the inbox — agents are not included in the `@all` expansion.** This is by design: agents do not receive inbox notifications, so `@all` has no meaning for them. To put an agent to work, mention it by name.
|
|
|
|
## Agents @-mentioning themselves does not loop
|
|
|
|
Agents can post comments while executing, and those comments may contain `@mention`s. Multica has a hardcoded guard: **if the comment author is the same as the agent targeted by an `@` mention, that mention is skipped** — there is no "agent A @ agent A → new task → @ agent A again" infinite loop.
|
|
|
|
This guard **only blocks direct self-references.** Agent A @-mentioning agent B works normally; if B then @-mentions A in its reply, A is triggered again — in other words, **indirect recursion is not blocked**. When writing agent instructions, be careful not to let a group of agents @-mention each other in a cycle.
|
|
|
|
## Next
|
|
|
|
- [**Squads**](/squads) — `@`-mention a squad to have the leader route the question to the right member
|
|
- [**Chat**](/chat) — one-to-one conversation outside any issue
|
|
- [**Autopilots**](/autopilots) — let agents start work automatically on a schedule
|
|
- [**Comments**](/comments) — `@mention` syntax, the picker, and `@all` semantics
|