eli5
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, including code, logs, and files, which serves as a potential surface for indirect prompt injection.
- Ingestion points: The skill explicitly instructs the agent to open and read code, files, logs, or errors provided in the user's request (SKILL.md, Step 2).
- Boundary markers: There are no explicit delimiters or sanitization steps mentioned to isolate user-provided content from the agent's instructions, though the skill does include a directive to "Stop after the explanation" and not perform other tasks.
- Capability inventory: The skill primarily uses file reading and text generation capabilities.
- Sanitization: No sanitization or filtering of the input content is specified.
Audit Metadata