Menu
Fixture exercising every obfuscation rule. Use only in tests.
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Basé sur la classification professionnelle SOC
Scans AI agent skill directories for dangerous bash patterns, prompt injection, supply chain risks, and SKILL.md structure violations. Use when reviewing, validating, or security-auditing a skill before deployment — or when asked to check a skill for security issues, prompt injection, hardcoded secrets, or compliance with Anthropic's authoring guidelines.
A skill with a Cyrillic а instead of a
Fixture exercising every PII rule. Use only in tests.
Fixture exercising every malicious-URL rule. Use only in tests.
A skill whose name does not match its directory. Use when testing the name-directory-mismatch rule.
Performs safe read-only filesystem operations. Use when the user needs to search, list, or inspect files without making changes.
| name | obfuscation-skill |
| description | Fixture exercising every obfuscation rule. Use only in tests. |
Plain prose should never trigger any rule even when it goes on for many words and many lines and never repeats.
A long base64 payload hidden in prose: VGhpc0lzU29tZUJhc2U2NEVuY29kZWREYXRhMTIzNDU2Nzg5MEFCQ0RFRkdISUpLTE0=
A long hex blob in prose: 0123456789abcdef0123456789abcdef0123456789abcdef0123456789abcdef
A high-entropy non-base64 token: aZ!Q9#7xL@2pN&8sK*4mR%6vT$3yU^1wE+0qO~5iY-jH=bG?cF/dD
Suppressed payload should be silent: VGhpc0lzU29tZUJhc2U2NEVuY29kZWREYXRhMTIzNDU2Nzg5MEFCQ0RFRkdISUpLTE0= # scan:ignore
# Inside a fenced code block — must be ignored:
VGhpc0lzU29tZUJhc2U2NEVuY29kZWREYXRhMTIzNDU2Nzg5MEFCQ0RFRkdISUpLTE0=
A URL is excluded by design: https://example.com/aZ9Q7xL2pN8sK4mR6vT3yU1wE0qO5iY8
Inline code should be ignored: VGhpc0lzU29tZUJhc2U2NEVuY29kZWREYXRhMTIzNDU2Nzg5MEFCQ0RFRkdISUpLTE0=