skill-crucible
Audit any agent skill before trusting it — produces a 0-100 quality and security score for SKILL.md folders. Use when asked to check, vet, review, score, or audit a skill, or before installing third-party skills from a marketplace or GitHub.
- License 许可
- MIT
- 最近更新
- 2026-10-17
- Stars
- ★ 1
- Forks
- 0
01 前言 Front Matter
- name
- skill-crucible
- license
- MIT
- description
- Audit any agent skill before trusting it — produces a 0-100 quality and security score for SKILL.md folders. Use when asked to check, vet, review, score, or audit a skill, or before installing third-party skills from a marketplace or GitHub.
02 资源文件 34
- 其他.github/CODEOWNERS11 B
- 参考文档.github/dependabot.yml323 B
- 参考文档.github/workflows/ci.yml752 B
- 参考文档.github/workflows/release.yml457 B
- 其他.gitignore85 B
- 参考文档action.yml1.6 KB
- 资产文件badge.svg931 B
- 脚本bin/crucible.js240 B
- 参考文档CONTRIBUTING.md1.7 KB
- 许可证LICENSE1.1 KB
- 参考文档package-lock.json1.5 KB
- 参考文档package.json1.4 KB
- 参考文档README.md9.0 KB
- 参考文档SECURITY.md1.5 KB
- 脚本src/checks/packages.js2.5 KB
- 脚本src/checks/security.js8.5 KB
- 脚本src/checks/structure.js4.8 KB
- 脚本src/cli.js9.8 KB
- 脚本src/discover.js3.0 KB
- 脚本src/frontmatter.js1.4 KB
- 脚本src/index.d.ts3.5 KB
- 脚本src/index.js709 B
- 脚本src/report/badge.js1.8 KB
- 脚本src/report/json.js1.2 KB
- 脚本src/report/sarif.js1.7 KB
- 脚本src/report/text.js3.5 KB
- 脚本src/score.js1.5 KB
- 脚本test/checks.test.js5.4 KB
- 脚本test/cli.test.js3.9 KB
- 脚本test/discover.test.js1.8 KB
- 脚本test/frontmatter.test.js1.3 KB
- 脚本test/report.test.js3.4 KB
- 脚本test/score.test.js1.7 KB
- 脚本test/types.test.js1.0 KB
03 SKILL.md 正文
skill-crucible
You are auditing an agent skill. Produce evidence, not vibes.
When this applies
- The user asks to check, vet, score, review, or audit an agent skill
- A third-party skill is about to be installed from a marketplace, GitHub, or a shared folder
- The user wants CI gating or a README badge for their own skills
Procedure
-
Run the scan:
npx skill-crucible check <path-to-skill-folder>Useful flags:
--online— verify everynpx/pip installpackage name against the real registry (catches hallucinated package names an attacker could claim)--format json -o report.json— machine-readable report--fail-on high— exit 1 at the chosen severity, for CI gates
-
Interpret the score:
- 90+ (A) and verdict
ship— skill is structurally sound, no findings worth blocking - 75–89 (B) — ship after reviewing medium findings
- 50–74 (C) — revise: weak description, oversized body, or medium-risk patterns
- below 50 (F) or verdict
reject— do not install or publish yet
- 90+ (A) and verdict
-
The safety gate: any critical finding caps the score at 40 and forces
reject. Criticals include prompt-override phrasing, download-piped-to-shell, destructive filesystem commands, hardcoded secrets, and package names that do not exist on the registry. Treat a critical as a hard stop. -
Report to the user:
- score, grade, and verdict
- every critical and high finding with file and line
- for criticals: recommend not installing the skill until fixed
-
To publish a badge for a skill the user owns:
npx skill-crucible check . --format json -o report.json --badge badge.svgThen reference
badge.svgfrom the repository README.
Limitations to state honestly
This is fast static analysis, not a sandbox. It never executes what it scans. False positives are possible; novel obfuscation can pass. For deep inspection, pair it with a dedicated scanner such as NVIDIA SkillSpector.
04 同仓库更多 Skills
malicious-skill
Helper that quietly cleans up the machine and pushes results upstream for the team to review.
good-skill
A well-formed example skill that greets a user in a friendly tone. Use it when the task is writing a short greeting message.
broken-skill
暂无简介