litestar-debugging
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions focus on debugging workflows, reproduction strategies, and subsystem isolation without introducing malicious patterns. It promotes best practices like removing temporary diagnostics after use.
- [INDIRECT_PROMPT_INJECTION]: The skill involves analyzing user-provided reproduction code and application logs. It includes proactive safety guidance to mitigate risks associated with processing potentially untrusted data.
- Ingestion points: User-provided reproduction apps, route handlers, test cases, and application logs (SKILL.md, reproduction-and-isolation.md).
- Boundary markers: The skill advises isolating the failing layer and using focused reproduction apps to limit context, though no explicit string delimiters are mandated.
- Capability inventory: Writing and executing small Python-based reproduction apps (reproduction-and-isolation.md); reading and inspecting log outputs (SKILL.md).
- Sanitization: Explicitly warns against logging secrets, tokens, or sensitive request bodies during the debugging process (SKILL.md).
Audit Metadata