skill-metadata-maintainer
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes content from repository files (SKILL.md and CONTRIBUTING.md) and Git metadata to automate documentation updates, which creates a potential surface for indirect prompt injection from repository contributors.\n
- Ingestion points: Reads local
SKILL.mdfiles and the project'sCONTRIBUTING.mdfile.\n - Boundary markers: Leverages YAML frontmatter formatting to isolate metadata from markdown content.\n
- Capability inventory: Contains file-write capabilities targeting local markdown files.\n
- Sanitization: Implements a mandatory human-in-the-loop confirmation workflow, requiring the user to review a diff of proposed changes before any write operation is performed.
Audit Metadata