[plan] Document obfuscation warnings as expected pattern in agentic workflows

[github-actions]

Objective

Create documentation explaining why dynamic command execution (obfuscation warnings from zizmor) is an expected and intentional pattern in gh-aw agentic workflows.

Context

From discussion #12276, 63 workflows generate 126 low-priority obfuscation warnings from zizmor security scanning. These warnings are expected because dynamic execution is a core feature of agentic workflows, not a security flaw.

Current Situation

• Tool: zizmor detects dynamic command execution as potential obfuscation
• Affected: 63 workflows (all agentic workflows with AI-generated commands)
• Status: Expected and acceptable ✅
• Mitigation: Runtime safety checks already in place

Required Documentation

Add a new documentation page explaining:

1. Why obfuscation is intentional

   • AI agents generate contextual commands at runtime
   • Enables flexible automation without hardcoded scripts
   • Core feature enabling adaptive workflow behavior

2. Safety measures in place

   • Sandboxed execution environment
   • Limited permissions scope (read-only for most agents)
   • Secret verification before execution
   • Audit logging of agent commands
   • Firewall checks on agent output

3. Trade-offs accepted

   • ✅ Enables agent flexibility
   • ⚠️ Reduces static auditability (accepted)
   • ✅ Mitigated by runtime monitoring

4. When to be concerned

   • Unexpected obfuscation in non-agentic workflows
   • Missing safety checks in agent execution
   • Overly permissive permissions

Suggested Location

• Create: docs/src/content/docs/security/dynamic-execution.md
• Update: docs/src/content/docs/security/index.md to reference new page

Acceptance Criteria

• Documentation clearly explains intentional dynamic execution pattern
• Safety measures are documented with examples
• Guidance provided for when obfuscation warnings should be investigated
• Linked from security documentation index
• Run make agent-finish before committing

AI generated by Plan Command for discussion #12276

[github-actions]

🍪 Issue Monster has assigned this to Copilot!

I've identified this issue as a good candidate for automated resolution and assigned it to the Copilot agent.

The Copilot agent will analyze the issue and create a pull request with the fix.

Om nom nom! 🍪

AI generated by Issue Monster