[plan] Add workflow complexity detection heuristics

[github-actions]

Objective

Implement heuristics to automatically detect workflow request complexity (basic/intermediate/advanced) based on user input characteristics.

Context

To support tiered complexity calibration (see related issue), we need algorithmic detection of request complexity. Research from #12193 shows uniform quality across diverse scenarios, indicating opportunity to differentiate response depth based on request sophistication.

Approach

1. Define complexity indicators:
   • Basic signals: Single trigger, standard GitHub events, no tool customization, simple English description
   • Intermediate signals: Multiple triggers, tool combinations (2-3 tools), conditional logic mentioned, specific file paths referenced
   • Advanced signals: Custom toolchains, performance requirements, multi-stage workflows, security customization, complex integration points
2. Implement scoring function in agent instructions or CLI
3. Set thresholds: Basic (0-3 points), Intermediate (4-7 points), Advanced (8+ points)
4. Pass detected tier to agent context for response calibration

Files to Modify

• .github/instructions/*.instructions.md - Add complexity detection logic
• pkg/cli/workflow_command.go or similar - Implement detection function
• Agent system prompts - Use detected tier in generation
• AGENTS.md - Document complexity detection algorithm

Acceptance Criteria

• Complexity detection function implemented
• Scoring algorithm defined with clear criteria
• Thresholds validated with test cases
• Detected tier passed to agent context
• Tests cover edge cases and boundary conditions
• Documentation explains detection methodology

Reference

Supports "Calibrate Complexity" recommendation from discussion #12193

AI generated by Plan Command for discussion #12193

[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