capability-audit
Skill53 starsChanged 52 days agoMITView on GitHub
Use when evaluating, comparing, installing, or recommending a Claude Code plugin, Agent Skill, MCP server, agent toolkit, or GitHub-hosted AI extension, especially when popularity is weak evidence or the candidate requests shell, network, filesystem, or credential access.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- snapsynapse/ai-tool-watchtheir own version, first seen Jul 23, 2026
- farky-tech/containerstheir own version, first seen Jul 13, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- snapsynapse/ai-tool-watch★ 11same name
- farky-tech/containers★ 1same 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 repository14 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- llms.txtllms.txt
- capability-auditSkill
- changelog-from-commitsSkill
- deterministic-checksSkill
- doc-link-auditSkill
- find-promptSkill
- skill-auditSkill
- changelog-from-commitsSkill
- deterministic-checksSkill
- doc-link-auditSkill
- find-promptSkill
- skill-auditSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 16
Run the script for all credible candidates, not only the popularity leader:
- Distinctive: line 22
Read the exact pinned commit or package that would be installed.
- Very specific: line 27
Use --json for machine-readable output and --token-env NAME only when an existing GitHub token environment variable is needed for rate …
- Very specific: line 41
Read ecosystem-discovery-guide.md for the source order, weighted scorecard, research seeds, and installation gate.
≈ “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.
Files
Capability Audit
Purpose
Collect comparable evidence about GitHub-hosted agent capabilities before installation. The bundled script uses the GitHub API without cloning a repository. It reports maintenance metadata, license, distribution files, executable surfaces, test signals, and high-risk text patterns. It does not install, authenticate, or declare a project safe.
Workflow
- Convert each candidate to
owner/repoor a GitHub repository URL. Run the script for all credible candidates, not only the popularity leader:
python ${CLAUDE_SKILL_DIR}/scripts/audit.py owner/repo another/repo- Read the exact pinned commit or package that would be installed. The script inspects the current default branch only; say so explicitly.
- Open every discovered plugin manifest,
SKILL.mdfrontmatter, hook, MCP configuration, and executable script before recommending installation. - Compare candidates using task fit, blast radius, runtime proof, maintenance, interoperability, rollback, documentation, and adoption. Cap adoption at 5% of the decision.
- Report four states separately: discovered, structurally inspected, installed/authenticated, and runtime-tested.
Use --json for machine-readable output and --token-env NAME only when an existing GitHub token environment variable is needed for rate limits. Never paste a token into the command.
Interpretation rules
- A clean report means only that no configured heuristic fired.
- Stars are context, not a quality score.
- Recent pushes can be automated noise; inspect the changed content.
- Missing
SECURITY.mdis a review signal, not proof of insecurity. - Marketplace inclusion and pinned SHAs improve provenance but do not establish runtime safety.
- Hooks, MCP servers,
allowed-tools, install scripts, and downloaded binaries define the effective trust boundary. - The official MCP reference servers are teaching implementations; evaluate production hardening independently.
Additional resource
Read ecosystem-discovery-guide.md for the source order, weighted scorecard, research seeds, and installation gate.
Remember
Inspect the artifact you will execute, at the version you will execute—not the reputation of its repository.
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.