specification-management

Specification management creates docs/specs directories with auto-incrementing IDs and README metadata. Initialize specs by feature name, then use spec.py tooling to generate TOML output and README decision logs that track progress through PRD, SDD, and PLAN phases.

The spec lifecycle orchestrates three phases: PRD (product requirements), SDD (system design), and PLAN (implementation planning). The Skill guides phase transitions with README decision tracking, auto-ID directory structure, and template resolution to maintain context and progress across each stage.

Yes. README decision logs embedded in each spec directory capture decisions, context, and progress milestones. The Skill integrates with spec.py to manage metadata in TOML format and maintain a decision audit trail throughout the spec lifecycle.

No. Auto-incrementing IDs are generated automatically during spec initialization, eliminating manual ID management. The Skill handles directory creation, TOML metadata output, and README generation, so you focus on spec content rather than file organization.

The Skill resolves templates from skills/template paths and applies them during spec initialization. Templates populate README structure, decision frameworks, and phase-specific guidance, reducing boilerplate and ensuring consistent spec formats across PRD, SDD, and PLAN documents.

Status checks read TOML metadata and README logs to report current phase, decision history, and completion markers. This confirms which stage each spec occupies in the PRD→SDD→PLAN workflow without manual file inspection.