/speckit.analyze
Agent skill wrapper for the Claude command /speckit.analyze.
When the original command text references {{INPUT}}, $1, or named arguments, map them from the user's current request.
Command Instructions
Read-only consistency check across all specification artifacts for a feature. No files are modified.
Steps
- Locate artifacts:
- If input names a feature or directory, use it. - Otherwise check .specify/specs/ and ask if multiple features exist. - Load all available: spec.md, plan.md, tasks.md, data-model.md, contracts/*.md, research.md. - Load .specify/memory/constitution.md.
- Run six consistency checks. Record each finding with: category, severity (CRITICAL / HIGH / MEDIUM / LOW), location, description.
- Constitution conflicts (auto-CRITICAL): Any plan or task that violates a MUST or SHOULD NOT principle. - Coverage gaps: Requirements with no corresponding tasks; tasks traceable to no requirement. - Inconsistency: Terminology drift, conflicting tech choices, ordering contradictions between artifacts. - Ambiguity: Vague terms (fast, scalable, secure) without measurable criteria; unresolved [NEEDS CLARIFICATION] markers. - Underspecification: Incomplete acceptance criteria, undefined component references, missing error conditions in contracts. - Duplication: Near-identical requirements or tasks appearing in multiple places.
- Produce analysis report in the conversation (do not write to any file):
## Consistency Analysis: <Feature Name> ### Findings (sorted CRITICAL → HIGH → MEDIUM → LOW, max 50 rows) | # | Severity | Category | Location | Finding | ### Coverage Matrix Each spec.md requirement → mapped tasks (or NONE) Each tasks.md task → source requirement (or NONE) ### Metrics - Requirement coverage: X% - Constitution violations: N - Ambiguous items: N - Duplicate items: N ### Recommended Next Actions - Offer remediation: If findings exist, ask: "Would you like me to fix any of these?" Apply only user-approved edits.
Next step: Address any CRITICAL findings, then run /speckit.tasks.
{{INPUT}}