🔒 Agentic Workflow Security Skill

Purpose

Comprehensive security guidance for GitHub Agentic Workflows implementing the official 5-layer defense-in-depth architecture: read-only tokens, zero secrets in agent, containerized execution with Agent Workflow Firewall (AWF), safe outputs with guardrails, and agentic threat detection.

When to Use

Apply this skill when:

• Designing secure agentic workflows
• Configuring safe-inputs, safe-outputs, and threat detection
• Setting up network firewall rules (AWF)
• Implementing integrity filtering for public repositories
• Responding to security incidents in agentic workflows

Rules

5-Layer Defense-in-Depth Architecture