defining-issues
Pass
Audited by Gen Agent Trust Hub on Feb 21, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- Indirect Prompt Injection (LOW): The skill transforms untrusted natural-language user requests into engineering documentation, which could allow malicious instructions to influence the agent's file-reading behavior or the content of generated issues.
- Ingestion points: User-supplied vague engineering requests.
- Boundary markers: Absent; there are no delimiters or specific instructions to disregard embedded commands in the input data.
- Capability inventory: The skill uses Glob and Read tools to inspect the codebase and writes new documentation to the docs/issues directory.
- Sanitization: Absent; the skill relies on the underlying model's safety constraints to handle potentially malicious input strings.
Audit Metadata