to-prototype

Pass

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions state that prototypes should be "trivial to run" using the project's task runner, explicitly mentioning commands like pnpm <name>, python <path>, or bun <path>. This encourages the execution of shell commands based on generated content.\n- [INDIRECT_PROMPT_INJECTION]: The skill generates executable code by ingesting user prompts and existing code context, which is a standard surface for indirect prompt injection.\n
  • Ingestion points: User prompts (e.g., "prototype this", "spike this") and existing project modules/pages.\n
  • Boundary markers: The skill does not provide instructions for using delimiters or boundary markers to isolate untrusted user input from the generation logic.\n
  • Capability inventory: The skill can write files (HTML, JS, TSX) and recommends command execution via shell task runners.\n
  • Sanitization: The instructions do not specify any sanitization or validation steps for input before it is used to generate the prototype code.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 8, 2026, 12:27 AM
Security Audit — agent-trust-hub — to-prototype