Upkeep audit

Skill wei18/Upkeep/skills/upkeep-audit

Upkeep — an AI audit crew for your repo. Catches docs/spec/asset drift with evidence; output-only. Claude Code plugin/skill + reusable CI workflow.

Install

npx -y skills add wei18/Upkeep --skill upkeep-audit

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

9 stars9 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Run the Upkeep repo audit locally against any repository path. Use when asked to audit a repo, check docs/spec/asset drift, or run upkeep without GitHub Actions.

SKILL.md

2.1 KB, as published. Nobody here has run it

Upkeep Local Audit

Run the full Upkeep audit pipeline (discovery → parallel reviewers → synthesis → HTML report) against a local repository. Output-only: the target repo is never modified.

Steps

Ensure the Upkeep checkout at ~/.cache/upkeep:
- Missing: git clone --depth 1 https://github.com/wei18/upkeep ~/.cache/upkeep
- Present: git -C ~/.cache/upkeep pull --ff-only (if the pull fails — diverged or dirty checkout — delete ~/.cache/upkeep and re-clone; it is disposable)
- Then (cd ~/.cache/upkeep && npm ci) after a fresh clone or whenever the pull changed package-lock.json (when unsure, run it — it is idempotent).
Run the audit (takes several minutes; reviewers run as parallel claude -p subprocesses — run it in the background and report progress):
```
~/.cache/upkeep/scripts/local-audit.sh <target-path> [--model M] [--rubric-lang L] [--max-turns N] [--out FILE]
```
Pass flags only when the user asked for them, with one exception: the report defaults to ./upkeep-report.html in the current working directory, so if the working directory is inside the target repo, pass --out pointing outside it (e.g. ~/upkeep-report.html) — the audit must never write into the target repo. Other defaults match the CI inputs (claude-opus-4-8, rubric en, 30 turns); --rubric-lang accepts en or zh-TW.
Summarize: the script prints the report markdown at the end. Present the findings grouped by severity (high → medium → low), each with its file path and a one-line problem statement, then give the absolute path of the generated report file (upkeep-report.html by default, or the --out path). If any reviewers failed, name them.

Requirements (tell the user what is missing instead of failing silently)

claude CLI installed and logged in (Claude Pro/Max subscription).
Node 20+, git.