auditor
Skill401 starsChanged 3 months agoApache-2.0View on GitHub
Run the domain-focused Auditor persona on the local working tree's diff against a base branch. May build/test if needed for confirmation. Outputs a verdict, optional suggested-changes patch, and (if relevant) a proposed PR description. Use after the Skeptic has cleared the branch, or directly when the user trusts their own code and wants the domain review.
We split this skill into the rules it gives an agent
5 rules. 0 are common, 2 are distinctive, and 3 are very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
2
Distinctive
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 24
If the diff is empty, report "No changes vs $BASE" and exit.
- Very specific: line 36
Spec version bump: if the diff touches runtime/ or pallets/ and spec_version in runtime/src/lib.rs was not bumped, do NOT modify the file.
- Very specific: line 38
Duplicate-work check: if a PR exists, run the same gh pr list check the persona file describes.
- Very specific: line 64
Write any suggested code changes to .auditor-suggestions.patch (apply with git apply).
- Distinctive: line 66
Do NOT post anything to GitHub.
“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Auditor — local mode
You are running the Auditor persona locally against the user's working tree. The Skeptic has either already passed (or the user is running you directly because they wrote the code themselves and trust intent). Your output goes to the terminal, not GitHub.
Step 1 — Determine the diff
Same detection as the Skeptic skill:
- PR base via
gh pr view --json baseRefNameif a PR exists. - Default to
devnet. - Override via skill argument:
/auditor main.
Compute the diff:
git fetch origin "$BASE" --quiet
git diff --merge-base "origin/$BASE"...HEADIf the diff is empty, report "No changes vs $BASE" and exit.
Step 2 — Run the persona
Load and follow:
.github/ai-review/common.md.github/ai-review/auditor.md
Local-mode adaptations:
- PR description handling: if a PR exists, follow the persona's auto-fill / discrepancy-comment logic but do NOT actually call
gh pr edit. Instead, write the proposed description to.auditor-pr-description.mdand tell the user. If no PR exists, generate a draft description and write it to the same file — the user will use it when they open the PR. - Auto-fix CI failures: you MAY run
./scripts/fix_rust.shagainst the working tree if lints / formatting are off, but DO NOT commit. Leave changes in the working tree for the user to review. - Spec version bump: if the diff touches
runtime/orpallets/andspec_versioninruntime/src/lib.rswas not bumped, do NOT modify the file. Instead, surface this as a finding the user must address. - Build/test escalation: same rules as the workflow — only build/test when a finding requires runtime confirmation. Use
cargo test -p <pallet> <test>for targeted tests rather than the full workspace. - Duplicate-work check: if a PR exists, run the same
gh pr listcheck the persona file describes. If no PR exists, skip this step (no duplicates to check yet).
Step 3 — Output
============================================================
AUDITOR VERDICT: 👍 | 👎
============================================================
Gittensor: KNOWN | LIKELY | UNKNOWN
Spec version: <bumped | NOT BUMPED — required>
Auto-fix: <ran fix_rust.sh, modified N files | not needed>
Description: <see .auditor-pr-description.md | already adequate>
Duplicates: <none | PR #N is the better candidate>
Findings:
[SEVERITY] Title
file:line — description
Suggested new files:
path/to/new_test.rs (see .auditor-suggestions.patch)
Conclusion: <one or two sentences>Write any suggested code changes to .auditor-suggestions.patch (apply with git apply). Write any proposed new files into the patch as well, as added-file diffs. Write the proposed PR description (if generated) to .auditor-pr-description.md.
Do NOT post anything to GitHub. Do NOT commit. Do NOT push.
Other versions5 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 5
- QinghongLin/data2story-skill★ 155same name
- rade-nugroho/Agentic-Enteprises-Skill★ 9same name
- daemon-blockint-tech/Agentic-Enteprises-Skill★ 8same name
- NotRedFox/NotRedFoxs-Claude-skills★ 1same name
- edisonzerolam/openclaw-skills★ 0same 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 repository5 other files it gives its agents.
- AGENTS.mdAGENTS.md
- clones/AGENTS.mdAGENTS.md
- .github/copilot-instructions.mdCopilot instructions
- evm-maintainerSkill
- skepticSkill
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
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.

