characterize
Pass
Audited by Gen Agent Trust Hub on Aug 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes project source code and subsystem entry points to generate tests, creating a vector where malicious instructions in the analyzed files could influence agent behavior. * Ingestion points: Project source code, entry points (CLI, functions, etc.), and user-provided subsystem arguments. * Boundary markers: Absent. * Capability inventory: Writing files (tests, fixtures) and executing local entry points (CLI, functions). * Sanitization: Absent.
- [DYNAMIC_EXECUTION]: The skill facilitates the execution of local code entry points to capture observable behavior for Golden Master snapshots. This behavior is the primary purpose of the skill but involves executing code within the local environment.
Audit Metadata