skill-manifest-generator
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill performs its intended function of auditing and documenting skill directories without exhibiting malicious patterns.
- [INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it reads and processes the contents of untrusted local files during manifest generation.
- Ingestion points: The
scripts/generate_manifest.pyscript reads the contents of all non-excluded files in a provided directory (e.g.,SKILL.md). - Boundary markers: No delimiters or specific instructions are used to ignore embedded instructions within the scanned files.
- Capability inventory: Includes file read access, SHA256 hashing, and writing results to a
MANIFEST.jsonfile. - Sanitization: Extracted data is written using
json.dump, which provides standard escaping for JSON strings.
Audit Metadata