prototype
Pass
Audited by Gen Agent Trust Hub on Jul 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to create and execute terminal scripts, test harnesses, and fixture runners to validate prototype behavior.
- [SAFE]: The instructions reduce risk by mandating that all work be isolated in specific disposable directories (e.g., docs/prototypes/) and prohibiting the modification of production code paths.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes user-provided requirements to generate and execute code. 1. Ingestion points: User input defining prototype requirements in SKILL.md. 2. Boundary markers: Absent. 3. Capability inventory: File system writes for prototype creation and subprocess execution for running code as described in SKILL.md. 4. Sanitization: Absent.
Audit Metadata