ai-evolution-engine
Pass
Audited by Gen Agent Trust Hub on Mar 18, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is designed to update core instruction files such as AGENTS.md and SOUL.md based on external learning data from sources like ClawHub. This creates an indirect prompt injection surface where malicious external content could alter the agent's fundamental logic. * Ingestion points: scripts/learn.mjs and scripts/evolve.mjs (referencing external scan results) * Boundary markers: None present in the logic * Capability inventory: File system writes to instruction files and shell script execution * Sanitization: No sanitization or validation of the 'learning' content is evident in the provided scripts.
- [COMMAND_EXECUTION]: The script 'scripts/assess.mjs' performs directory traversal by reading the parent directory ('../../skills') to inventory other installed skills. Accessing files outside the skill's own directory scope is a potential risk factor for data exposure.
Audit Metadata