agentleFS
Sign inStart free
ilya-kozyrevilya-kozyrev/claude-agent-hub/skills/agy-review/SKILL.md

agy-review

Skill5 starsChanged 39 days agoMITView on GitHub

Review a branch with Gemini 4 Argon through the Antigravity CLI (agy), detached and read-only. A reviewer skill for the hub (the contract in docs/reviewers.md), usually as an optional second opinion from another vendor. Use it when `hub reviewer` prints "load skill `delamain:agy-review`"; it takes a brief file, a checkout, a base sha and a head ref.

2 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions2 projects use a version of this.

    Word for word · 1

    1. ilya-kozyrev/delamain★ 5100% same

    Edited · 0

    None found.

    Wrote their own version · 1

    1. OctavianTocan/agy-review★ 2same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository8 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    The hub verifies every finding against the code: before it goes to the author: a review is a colleague's opinion.

    line 23

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Argon review through agy

Inputs from the hub: BRIEF (a file, from templates/brief-review.md), REPO (a checkout that has the head commit), BASE (a sha), HEAD (a ref; the brief names git diff BASE..HEAD). The tool is agy-review in the plugin's bin/ (on PATH under Claude Code; otherwise <plugin-root>/bin/agy-review, the root being PLUGIN_ROOT in Codex and CLAUDE_PLUGIN_ROOT in Claude). The review itself is Argon's, run by agy; this skill only starts it and says how to wait.

  1. Check. agy-review check: exit 0 means agy is installed and Argon's weekly quota is not known to be below 20 %. Anything else: say so and stop; the hub takes the next reviewer. (run repeats the check and refuses with exit 3.)
  2. Start. agy-review run --cwd REPO --brief BRIEF --name review-<branch> [--add-dir DIR]... [--stage S]. It prints run: <dir> and model: <id> and returns at once; the review runs detached. Material outside REPO (the author's report, the spec) goes in with --add-dir, because the reviewer reads the checkout and those directories only.
  3. Wait for the file <dir>/rc with one jwait --file <dir>/rc, in the background; no polling loop. With --stage the run also journals a start line and REVIEWED <dir>/final.md — rc 0 (or EXIT agy-review …).
  4. Result. rc is 0: the review is <dir>/final.md; say that path in your final message. 2: agy ended without an answer even after two resumes: read out-N.json and stderr.log, say what happened, and the hub decides. Any other value is agy's own exit code (128+N after signal N).
  5. The hub verifies every finding against the code before it goes to the author: a review is a colleague's opinion.

Cost and quota. One review of a real change costs 6–13 % of the Google AI Pro weekly Argon quota and takes 6–12 minutes (agy-review limits shows what is left; --min-left PCT or AGENT_HUB_AGY_REVIEW_MIN_LEFT moves the floor of 20 %). That is why the config entry marks it "second": true and the hub asks for it only when a miss would be costly. In a trial on two real merge requests Argon and Codex found about as many real defects, and different ones.

Read-only, in three lines. The review runs under its own HOME (<hub home>/agy-review-home) with a settings.json the tool rewrites on every run: only commands that cannot start another program are allowed (git diff/log/show, grep, cat, head, tail, ls, …); writes, the web, MCP, shells, interpreters, awk, find, sed, rg, sort, xargs, build tools and network and database clients are denied. agy runs with --sandbox, which leaves the network open, so it is the list that keeps programs from starting; --dangerously-skip-permissions is never used. Your own ~/.gemini/antigravity-cli/settings.json is neither read nor written.

Needs git 2.38 or newer on PATH (older is refused). Verified on macOS with agy 1.4.0 only. Linux and Windows are untested.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free