figure-it-out
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious instructions, obfuscation, or unauthorized access patterns were found. The skill provides procedural guidelines for managing complex development tasks and maintaining a decision trail.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-defined tasks and artifacts, which is an inherent part of its planning function. It includes safety-oriented features like verification against predicates and audit logs. (1) Ingestion points: User-provided task descriptions and artifacts from delegation. (2) Boundary markers: Usage of falsifiable predicates to ground definitions of done. (3) Capability inventory: File writing for audit logs, script execution for verification, and delegation to worker agents. (4) Sanitization: Not explicitly specified in instructions.
Audit Metadata