prototype
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill establishes strong architectural boundaries for prototypes, such as using in-memory state and separate branches, which minimizes production impact.
- [COMMAND_EXECUTION]: The skill's workflow expects prototypes to be runnable via standard project task runners (e.g., pnpm, python, bun). This capability is intended for testing generated prototypes and does not instruct the agent to perform unauthorized command execution.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it ingests untrusted user prompts to generate executable prototypes. 1. Ingestion points: User prompts and project source code (e.g., in SKILL.md and LOGIC.md). 2. Boundary markers: Not explicitly required in the instructions. 3. Capability inventory: File writing (HTML, JS, TSX) and command execution (pnpm, python, bun). 4. Sanitization: No specific input sanitization or validation rules are defined for the code generation process.
Audit Metadata