md-review
Pass
Audited by Gen Agent Trust Hub on Aug 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted Markdown files via the --input argument across its analysis scripts, creating an attack surface for indirect prompt injection. Maliciously crafted documentation could include hidden instructions aimed at the agent. However, the skill's capabilities are confined to static analysis and terminal reporting, significantly limiting potential impact.
- Ingestion points: File content is read in scripts/md_review_gate.py, scripts/link_checker.py, and scripts/readability_scorer.py.
- Boundary markers: Absent.
- Capability inventory: The scripts are restricted to local file read operations and printing findings to stdout; no network access or write operations are implemented in scripts/md_review_gate.py, scripts/link_checker.py, or scripts/readability_scorer.py.
- Sanitization: Regular expressions are used to isolate prose from structure, but do not provide security sanitization against prompt injection.
- [SAFE]: The skill explicitly avoids network requests by design, as confirmed by the implementation in scripts/link_checker.py which inventories external URLs without fetching them. No sensitive file paths, hardcoded credentials, or dangerous command execution patterns were detected in the source code.
Audit Metadata