log-debug-issue
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
- [NO_CODE] (SAFE): The skill consists entirely of markdown instructions and contains no executable scripts or binary files.
- [PROMPT_INJECTION] (LOW): Indirect Prompt Injection Surface (Category 8).
- Ingestion points: Processes error messages, stack traces, and environment details from the debugging context in SKILL.md.
- Boundary markers: Relies on markdown structure; lacks explicit delimiters to prevent the agent from interpreting instructions contained within logs.
- Capability inventory: Limited to file system write operations for the DEBUG_LOG.md file.
- Sanitization: Does not sanitize or escape external content before interpolation.
Audit Metadata