aisync is a local CLI for Multi-Agent Context Management, Coordination & Observability across AI coding tools. I maintain it. Two agents can write to its journal with aisync log, and aisync status shows who is blocked. I tested how blockers are closed and what --project defaults to, and found that one unrelated unblocked event can clear every open blocker in a project.
I used a throwaway hub (AISYNC_HOME and HOME in a temp directory) and aisync-cli 0.2.11 from PyPI, run with uvx.
An unrelated unblocked closes everything
Agents A and B each log a blocker in project shop2, and aisync status lists both:
aisync log -a A -p shop2 -k blocked -m "A: waiting for the API key" aisync log -a B -p shop2 -k blocked -m "B: staging database is down" aisync status BLOCKED A shop2: A: waiting for the API key (since 11-16 08:59) BLOCKED B shop2: B: staging database is down (since 11-16 08:59) Agent C then logs a message that has nothing to do with either blocker:
aisync log -a C -p shop2 -k unblocked -m "C: finished my own unrelated task" aisync status Both BLOCKED lines are gone. The rule is in ai_sync/core.py (build_current): an unblocked event closes the blocker whose id is in its blocked_id ref, and when there is no blocked_id it closes every open blocker with the same project. Who logged it does not matter.
The precise form works. aisync log prints the id of the event it wrote (logged a7acf21bd4ad35da). Passing it back closes only that blocker:
aisync log -a A -p shop3 -k unblocked --ref blocked_id=<id of A's blocked event> -m "A: key arrived" In a second project, with A's and B's blockers open, this left only B's BLOCKED line.
What I did not find is any documentation for it. blocked_id appears in core.py, board.py and the built-in demo, but not in the README, in the docs folder, in the --ref help (key=value, e.g. pr=45), or in the MCP text. The MCP log tool is declared as log(agent, project, summary, kind, files), with no way to pass a ref. An agent that reaches aisync through MCP can only write the broad form.
--project defaults to the checkout's directory name
The help says default: name of the current git repo. In practice it is the name of the directory that git rev-parse --show-toplevel returns:
Where aisync log ran | project recorded |
|---|---|
repo root myrepo | myrepo |
myrepo/src/deep | myrepo |
| a plain directory, no git | - (the event is still written) |
a git worktree directory wt-feature of myrepo | wt-feature |
Claims compare project names for equality, so a claim from the worktree does not meet a claim from the main checkout. A claimed src/a.py with the default project in myrepo. Then B claimed the same path from wt-feature with exit 0 (claimed wt-feature [src/a.py] for B), while C in myrepo got exit 3. Two worktrees are separate checkouts, so this may be what you want. If you use worktrees to give each agent its own branch and want claims to span them, pass the same --project everywhere. Outside a git repo, aisync claim without -p stops with error: claim needs --agent, --project and --summary (exit 2).
aisync log -k claim is only a journal line
claim and release are listed among the kinds log accepts. Logging one does not claim anything:
aisync log -a A -p myrepo -k claim -f docs/z.md -m "claimed docs/z.md" aisync claims no active claims B then claimed docs/z.md with exit 0. Use aisync claim to claim.
What this does not show
- One version (0.2.11), one machine (macOS, arm64), the CLI only. For the MCP
logtool I read the source and did not run it. - I tried one git worktree layout. I did not test submodules.
- I changed nothing in the code. Each finding is filed with the commands above: the broad
unblockedis https://github.com/MaxonT/aisync/issues/34, the worktree project name is https://github.com/MaxonT/aisync/issues/35, andlog -k claimis https://github.com/MaxonT/aisync/issues/36. Whether anunblockedwithout an id should close one blocker, the latest one by that agent, or all of them is a decision for the maintainer.
If you build something like this
Make closing a blocker name the blocker, or at least restrict it to the agent that opened it. Put the id in the output of the command that opens it, as aisync does, and document that it exists.
- Repo (MIT): https://github.com/MaxonT/aisync
- Site: https://maxont.github.io/aisync/
- Install:
pip install aisync-cli
This post and the test runs were written and run by an AI agent (Claude Code) for the maintainer. The outputs above are copied from those runs.