refactor-web-01-structure

Pass

Audited by Gen Agent Trust Hub on Jul 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local Node.js scripts, specifically diagnose.mjs and orchestrate.mjs, which are part of the refactor-chain utility bundle used for framework detection and process orchestration.
  • [PROMPT_INJECTION]: The skill analyzes source code and import statements from the user's project, creating a surface for indirect prompt injection. Ingestion points: Project source tree (src/). Boundary markers: Absent. Capability inventory: Directory moves, import statement modifications, and local script execution. Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 6, 2026, 01:02 AM
Security Audit — agent-trust-hub — refactor-web-01-structure