business-logic
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted application code for analysis, which presents a surface for indirect prompt injection. 1. Ingestion points: Source code files specified via the scope flag. 2. Boundary markers: Not present. 3. Capability inventory: Manual analysis and semgrep scanner invocation. 4. Sanitization: Not specified for audited content.
- [SAFE]: The skill's operations, including the use of semgrep for business logic detection and the provided reference patterns, are consistent with legitimate security research and auditing. No evidence of credential harvesting, unauthorized network communication, or code obfuscation was found.
Audit Metadata