code-of-conduct-generator
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONMETADATA_POISONING
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest user requests to generate code and documentation but lacks explicit safety instructions or delimiters to protect against malicious instructions embedded in that input.
- Ingestion points: Processes user input triggered by keywords related to code and documentation in SKILL.md.
- Boundary markers: No delimiters (e.g., XML tags or triple quotes) or 'ignore embedded instructions' warnings are present to isolate user data from the agent's instructions.
- Capability inventory: The skill possesses extensive system capabilities, including the
Write,Edit, andBash(cmd:*)tools, which could be exploited if the agent is tricked by malicious user input. - Sanitization: The instructions do not define any validation, filtering, or escaping mechanisms for external content.
- [METADATA_POISONING]: The skill uses an overly broad trigger in its description that may cause it to activate for unrelated tasks.
- Evidence: The description field in SKILL.md specifies that the skill should trigger on generic phrases like "code generator" or simply "code", which could lead to the agent switching to this specific documentation context for any general programming request.
Audit Metadata