doc-standard
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill performs document validation by executing included Python scripts via the shell.
- Evidence: SKILL.md and README.md describe the usage of
scripts/check_headings.py,scripts/check_vocabulary.py, andscripts/check_traceability.pyfor deterministic audits. - [INDIRECT_PROMPT_INJECTION]: The skill analyzes user-provided Markdown files, representing a potential surface for indirect prompt injection through external data ingestion.
- Ingestion points: User-supplied markdown documents passed to checking scripts.
- Boundary markers: The agent uses pre-defined rubric files (e.g.,
rubric/srs.md) to guide qualitative assessments. - Capability inventory: The agent reads files and executes local scripts but does not possess network or broad filesystem write capabilities.
- [SAFE]: No malicious patterns such as credential theft, data exfiltration, or obfuscated payloads were detected. The author context provided for 'marycache' is consistent with the resources found, and the skill's behavior aligns with its documentation.
Audit Metadata