afk-pipeline

Pass

Audited by Gen Agent Trust Hub on Aug 12, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches and installs standard development tools such as Bun, GitHub CLI, Claude Code, Cursor, and Grok from their official, well-known domains during the sandbox build process.
  • [REMOTE_CODE_EXECUTION]: Executes official installation scripts for well-known developer tools using piped shell commands. These downloads originate from trusted providers and are documented neutrally.
  • [COMMAND_EXECUTION]: Runs a multi-phase implementation and verification loop that includes a security integrity check. The 'assertControlPlane' function compares the active orchestration code against a trusted git snapshot to ensure the pipeline has not been tampered with by implementing agents.
  • [CREDENTIALS_UNSAFE]: Manages necessary repository secrets and provider authentication tokens to enable automated code implementation and PR creation. These are handled using standard CI/CD practices and passed to isolated sandbox environments.
  • [PROMPT_INJECTION]: Ingests untrusted data from GitHub issues but mitigates potential injection attacks through a dedicated neutralization module. The 'neutralizeHostileText' function sanitizes content by stripping HTML, neutralizing URLs, and preventing unwanted mentions before generating PR comments.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 12, 2026, 07:49 PM
Security Audit — agent-trust-hub — afk-pipeline