What problem does it solve? In multi-phase physics projects, conventions like metric signature, Fourier transform definitions, and unit systems can silently drift between phases, invalidating downstream results. This Skill detects sign errors, notation drift, and convention lock violations before they propagate. ## Core Features & Use Cases - Convention Lock Drift Detection: Compares the machine-readable convention lock in state.json against the human-readable CONVENTIONS.md and flags disagreements as critical issues. - Cross-Phase Consistency Scanning: Extracts convention declarations from every completed phase SUMMARY.md and cross-references them against the project-level lock, classifying findings as CRITICAL, WARNING, or INFO. - Automated Conflict Resolution: Spawns consistency-checker and notation-coordinator subagents to validate conventions deeply and resolve critical conflicts with conversion tables. - Use Case: After completing several phases of a quantum field theory project, run this Skill to confirm that the metric signature and Fourier convention used in Phase 3 match those locked in Phase 1, catching a sign flip before it corrupts later calculations. ## Quick Start Ask the AI to validate conventions across all phases, or pass a phase number such as 3 to limit the validation scope to that phase only.