Checklists

requirements.md

Specification Quality Checklist: Doctrine Catfooding

Purpose: Validate specification completeness and quality before proceeding to planning Created: 2026-07-01 Feature: spec.md

Content Quality

  • ✅ No implementation details beyond what the domain requires — Note 1: doctrine-authoring mission; artifact kinds + existing-target file paths are the domain, not incidental tech
  • ✅ Focused on user value and business needs (catfooding: enforceable governance for contributors + consumers)
  • ✅ Written for stakeholders — Purpose + Background legible; the reconciliation table is load-bearing
  • ✅ All mandatory sections completed

Requirement Completeness

  • ✅ No [NEEDS CLARIFICATION] markers remain
  • ✅ Requirements are testable and unambiguous
  • ✅ Requirement types separated (FR / NFR / C)
  • ✅ IDs unique across FR-###, NFR-###, C-###
  • ✅ All requirement rows have a non-empty Status (all Draft)
  • ✅ Non-functional requirements include measurable thresholds (0 skipped artifacts; guard/lint green; non-shallow closure)
  • ✅ Success criteria measurable
  • ✅ Success criteria technology-agnostic — outcome-framed; verification names doctrine surfaces (Note 1)
  • ✅ Acceptance scenarios defined (primary catfooding loop + exceptions A/B/C)
  • ✅ Edge cases identified (duplicate-authority, required-directive contradiction, capstone order)
  • ✅ Scope clearly bounded (In/Out of scope)
  • ✅ Dependencies and assumptions identified

Feature Readiness

  • ✅ All FRs have clear acceptance criteria (via SCs + the per-conversion DoD triad C-002)
  • ✅ User scenarios cover primary flows
  • ✅ Meets measurable outcomes in Success Criteria
  • ✅ No implementation details leak beyond the domain (Note 1)

Notes

  • Note 1 (doctrine-authoring domain): This mission's product is doctrine artifacts + the compiled charter, so the spec necessarily names artifact kinds, specific existing extension targets (DIRECTIVE_041, etc.), and the charter machinery. This is load-bearing, not implementation leakage: the whole review finding is that the wrong artifact kind / a duplicate authority is the failure mode. Requirements remain outcome-framed (a rule is reconciled, the charter compiles, doctor is green).
  • All items pass. Ready for /spec-kitty.plan. The review artefacts in research/ are the authoritative inputs for planning.