ralph-loop
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill's primary function is to provide a framework for high-quality software development, focusing on project management and verification without malicious intent.\n- [COMMAND_EXECUTION]: The methodology involves running standard development commands for building and testing code.\n
- Evidence: Mentions executing
npm test,npm run build, andcurlfor verifying the correctness of implemented features within the developer sandbox.\n- [EXTERNAL_DOWNLOADS]: The skill references the use of standard package managers to handle project dependencies.\n - Evidence: Mentions
npmandpipinstallation commands within the context of project setup and automated quality checks.\n- [PROMPT_INJECTION]: The skill includes defensive instructions designed to maintain agent focus and prevent the implementation of unrequested or unsafe features.\n - Evidence: Defines 'Red Flags' (e.g., forbidding 'extra' features) and '3-strike' debugging rules to ensure the agent remains within the bounds of defined specifications and handles errors systematically.
Audit Metadata