ralph-wiggum-v2
Pass
Audited by Gen Agent Trust Hub on Jun 14, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from the local codebase being analyzed.
- Ingestion points: The skill reads source code and configuration files from the project directory (e.g., in the 'Discovery & Initialization' and 'Parallel Agent Review Swarm' phases).
- Boundary markers: The instructions do not define clear boundaries or 'ignore' directives to prevent the agent from following instructions accidentally embedded in code comments, documentation, or string literals within the project.
- Capability inventory: The skill has the capability to write and modify local files, execute shell commands (test suites, build scripts), and spawn additional agent tasks.
- Sanitization: No sanitization or validation is performed on the codebase content before it is processed by the review agents.
- [COMMAND_EXECUTION]: The skill's workflow involves the autonomous execution of code and shell commands throughout the TDD cycle.
- Evidence: The 'Iteration Loop' and 'Final Verification' sections describe an automated process for running full test suites, TypeScript checks, and production build commands ('Build production bundle') without per-step user confirmation.
Audit Metadata