refactor-ci-agent

Pass

Audited by Gen Agent Trust Hub on Jul 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [SAFE]: The skill promotes security-first configurations by providing 'Before vs After' examples that demonstrate how to avoid credential leakage and mitigate supply chain risks through action pinning.
  • [COMMAND_EXECUTION]: The skill documents the execution of a CI pipeline that runs diagnostic scripts (diagnose.mjs, guidelines.mjs, orchestrate.mjs) and an optional, user-defined agent command (agent-cmd). The documentation explicitly warns against including credentials in the command string.
  • [EXTERNAL_DOWNLOADS]: The skill references standard GitHub Actions, including actions/checkout and actions/setup-node, as well as the vendor's own composite action for generating reports.
  • [SAFE]: No malicious patterns, prompt injections, or obfuscated code segments were detected. The included utility script (scripts/checklist.mjs) is benign and used only for formatting internal checklists.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 6, 2026, 01:01 AM
Security Audit — agent-trust-hub — refactor-ci-agent