[Feature Request] Spec-to-Implementation Gap Classification and Metrics

[kanfil]

Summary

Add gap classification (Intent→Spec vs Spec→Implementation) and track spec quality metrics for harness improvement.

Problem

InfoQ article "Long-Term Direction" discusses:

• Classifying gaps as "Intent→Spec" or "Spec→Implementation"
• Using gap data to improve specification harness
• Tracking spec quality as governance metrics

Current State

✅ Already Implemented:

• /spec.trace - Session trace generation

  • Templates: templates/commands/trace.md, templates/trace-template.md
  • Scripts: scripts/bash/generate-trace.sh, scripts/powershell/generate-trace.ps1
  • Scripts: scripts/bash/validate-trace.sh, scripts/powershell/validate-trace.ps1
  • Tests: tests/test_trace_command.py
  • 5-section trace: Session Overview, Decision Patterns, Execution Context, Reusable Patterns, Evidence Links

• /spec.analyze - Cross-artifact validation

  • Detects discrepancies between spec, clarifications, plan, tasks, constitution
  • Non-destructive alignment report

❌ Still Needed:

• Gap classification (Intent→Spec vs Spec→Implementation)
• Metrics tracking (quality scores, gap frequency)
• Feedback to constitution improvement

Proposal

1. Gap Classification

• In /spec.analyze, classify gaps by type:
  • "Intent→Spec": Missing requirements, unclear acceptance criteria
  • "Spec→Implementation": Code doesn't match spec, incomplete features

2. Metrics Tracking

• Track: spec quality score, gap frequency, revision cycles
• Dashboard or report showing spec health over time

3. Feedback to Constitution

• Use gap data to suggest constitution improvements
• Pattern: repeated gaps → constitution rule addition

Priority

Low-Medium - Advanced feature, differentiates from competitors