gentle-ai-skill-registry
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONPERSISTENCE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted metadata (name and description) from all files matching
*/SKILL.mdin known directories to generate a centralized index. - Ingestion points: Frontmatter of scanned
SKILL.mdfiles (Step 2). - Boundary markers: No specific delimiters or "ignore instructions" markers are defined for the output registry file.
- Capability inventory: Writes to the local filesystem (
.atl/skill-registry.md), modifies.gitignore, and persists data to Engram memory. - Sanitization: No sanitization or validation of the extracted name and description triggers is performed before they are indexed.
- [PERSISTENCE]: The skill includes instructions to automatically modify the project's
.gitignorefile to ensure the registry directory is persistent across sessions while being hidden from version control.
Audit Metadata