The Security Architecture of GitHub Agentic Workflow

GitHub extended GitHub Actions with agentic workflows that deploy AI agents to automate code tasks like documentation fixes, testing, and refactoring. Rather than assuming agents are trustworthy, GitHub's security architecture treats agents as potentially compromised, using isolation, restricted permissions, and input validation to prevent prompt injection attacks and unauthorized access to secrets.