human-readable-code
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFENO_CODEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill consists exclusively of markdown documentation and YAML configuration. No executable code (e.g., JavaScript, Python, Bash) is bundled with the skill.
- [EXTERNAL_DOWNLOADS]: The skill includes references to various external documentation sources (e.g., C++ Core Guidelines, PEP 8, Google Engineering Practices). These are informational links to well-known and reputable technical resources and do not represent an automated remote code execution risk.
- [INDIRECT_PROMPT_INJECTION]: The skill directs the agent to read project code and instructions to perform its task.
- Ingestion points: SKILL.md refers to reading request, project instructions, nearby code, callers, and tests.
- Boundary markers: None present; the skill relies on the agent's native handling of data vs. instructions.
- Capability inventory: The skill itself contains no executable scripts or system calls. The agent using the skill may have file-read/write capabilities as part of its core persona.
- Sanitization: No specific sanitization or filtering logic is prescribed for the ingested content, which is standard for a code-assistance guideline.
Audit Metadata