document-audit-extraction
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill performs local file system operations to inventory and analyze document structure. It uses
pathlibfor file discovery andyaml.safe_loadfor metadata extraction, which is a secure practice for parsing untrusted YAML content.\n- [INDIREC T_PRO MPT_INJE CTI ON]: The skill ingests untrusted data from local files to perform structural analysis.\n - Ingestion points: File reading via
Path.read_text()ininventory_documentswithinSKIL L.md.\n - Boundary markers: None explicitly defined for the script output processed by the agent.\n
- Capability inventory: File system read access, metadata extraction, and regex-based feature extraction.\n
- Sanitization: Implements
yaml.safe_load()to mitigate risks during frontmatter parsing.
Audit Metadata