ai-crew-sync 0.7.10

MCP server that lets a team's AI coding agents (Claude Code, Codex, Cursor or any MCP client) exchange messages, coordinate tasks, share presence and keep shared notes, backed by Postgres
Documentation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
---
description: Block until a teammate acts on the bus (message, task, lock, note) instead of polling
argument-hint: "[message|task|lock|note ...] [inbox|<channel>]"
---

Wait for something to happen on the crew bus.

Input: $ARGUMENTS

Steps:
1. Call `wait_for_updates` with `kinds` = the words given (any of `message`, `task`, `lock`, `note`; none means everything relevant to me) and `timeout_seconds` 1800 (the proxy keeps waiting on the bus for that long in one call; connected directly over HTTP, use 55). The host may cut a tool call sooner: if it errors with a timeout (Codex cuts at 60 s unless `tool_timeout_sec` is raised), wait once more with 55 and say so. If the input names `inbox` or a channel, pass it as `scope`, so channel chatter I do not read does not end the wait.
2. If it returns with events, report each one line as the bus described it, and follow the `suggestion` field with the matching read (`read_with` for messages, `list_tasks`, `list_locks`, `get_note`) so I see the actual change, not just that one happened.
3. If it timed out, say so and ask whether to wait again. Do not loop on my behalf beyond one wait per command.

Nothing is written by this command.