Audit Library Health
Runs ai-agent-skills doctor, check and curate review to audit a skills library workspace
Test report
- Verdict
- Works with setup
- Score
- Tested
- Jul 21, 2026
- Environment
- Claude Code 2.x (agent harness)
- Upstream re-checked
- Aug 10, 2026 · 5454cd3
Installed ai-agent-skills@4.3.2 from npm and ran all five workflow steps for real, first in a bare dir and then inside a workspace created with `init-library`. Three steps work as written (`doctor --format json` flagged 3 broken entries in my ~/.claude/skills; `check --format json` scanned 249 installs; `curate review` printed an empty queue), but two do not: bare `npx ai-agent-skills validate` at the library root returns status "error / No SKILL.md found" because the shipped `validate` is a per-skill-directory validator, not the catalog check the body describes, and `build-docs --dry-run --format json` dies with ENOENT on `node_modules/ai-agent-skills/WORK_AREAS.md` (cli.js passes the workspace context into `generatedDocsAreInSync`'s options slot, so it falls back to the bundled package dir, and WORK_AREAS.md is excluded from package.json `files`) — so the `currentlyInSync` value the skill tells you to branch on is unreachable, while plain `build-docs` succeeds. Baseline comparison: `npx ai-agent-skills help` already lists doctor/check/validate/build-docs/curate review plus `--format json` and `--dry-run`, and describes validate correctly, so an agent with no skill produced a 5-for-5 clean audit while the skill body produced one false failure and one crash; the Gotchas and Health Checklist sections are accurate and are the skill's only real added value. No security smells found — no curl|sh, no base64 blobs, no secret exfiltration, no injection text; `check` does make outbound GitHub requests, which the skill discloses.
Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works
- Installs cleanly 4/5
- Triggers reliably 5/5
- Output vs. baseline 3/10
- Docs & honesty 2/5
What Audit Library Health does
A five-step checklist for auditing an ai-agent-skills library workspace: validate the catalog, run doctor on installed skills, check upstream sources for updates, confirm generated README/WORK_AREAS docs are in sync, and review the curation queue. Triggers when you ask whether a skills library is healthy, consistent, or safe to push to a shared repo. It is a thin wrapper over the `npx ai-agent-skills` CLI and does nothing without that CLI and an initialized workspace.
How to install Audit Library Health
mkdir -p ~/.claude/skills/audit-library-health
curl -sL https://raw.githubusercontent.com/MoizIbnYousaf/Ai-Agent-Skills/HEAD/skills/audit-library-health/SKILL.md \
-o ~/.claude/skills/audit-library-health/SKILL.md
# Required runtime — the skill is only CLI invocations:
npm i -g ai-agent-skills # or use npx per-command
# Steps 1/4/5 need an initialized workspace, not just an install:
npx ai-agent-skills init-library my-library
cd my-library
Skills live in ~/.claude/skills/ (global) or .claude/skills/
(per-project). Restart Claude Code after installing.
Commands — how to trigger Audit Library Health
-
/audit-library-healthRuns ai-agent-skills doctor, check and curate review to audit a skills library workspace
It also activates on plain-language prompts like these:
-
Check the health of our skills library -
Validate all skills and flag stale ones -
Verify our generated docs are in sync
Frequently asked questions
- Is the Audit Library Health skill free?
- Yes. The skill itself is free from MoizIbnYousaf/Ai-Agent-Skills. SkillProof publishes the install command and an independent test verdict at no cost.
- Does Audit Library Health work with Claude Code?
- We tested it with Claude Code 2.x (agent harness) on Jul 21, 2026. Verdict: Works with setup. Installed ai-agent-skills@4.3.2 from npm and ran all five workflow steps for real, first in a bare dir and then inside a workspace created with `init-library`. Three steps work as written (`doctor --format json` flagged 3 broken entries in my ~/.claude/skills; `check --format json` scanned 249 installs; `curate review` printed an empty queue), but two do not: bare `npx ai-agent-skills validate` at the library root returns status "error / No SKILL.md found" because the shipped `validate` is a per-skill-directory validator, not the catalog check the body describes, and `build-docs --dry-run --format json` dies with ENOENT on `node_modules/ai-agent-skills/WORK_AREAS.md` (cli.js passes the workspace context into `generatedDocsAreInSync`'s options slot, so it falls back to the bundled package dir, and WORK_AREAS.md is excluded from package.json `files`) — so the `currentlyInSync` value the skill tells you to branch on is unreachable, while plain `build-docs` succeeds. Baseline comparison: `npx ai-agent-skills help` already lists doctor/check/validate/build-docs/curate review plus `--format json` and `--dry-run`, and describes validate correctly, so an agent with no skill produced a 5-for-5 clean audit while the skill body produced one false failure and one crash; the Gotchas and Health Checklist sections are accurate and are the skill's only real added value. No security smells found — no curl|sh, no base64 blobs, no secret exfiltration, no injection text; `check` does make outbound GitHub requests, which the skill discloses.
- What is the Audit Library Health SkillProof Score?
- 5.6/10 — installs cleanly 4/5, triggers reliably 5/5, output vs. baseline 3/10, docs & honesty 2/5.
- How do I install Audit Library Health?
- Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
- Can I use Audit Library Health with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
- The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.