cmd-speckit-checklist（cmd speckit 清单）

/speckit.checklist

Agent skill wrapper for the Claude command /speckit.checklist.

When the original command text references {{INPUT}}, $1, or named arguments, map them from the user's current request.

Command Instructions

Generate requirements quality checklists — "unit tests for requirements writing." These validate that specs are complete, clear, and unambiguous. They do NOT test implementation behavior.

Steps

1. Parse input: From the provided input, identify a quality domain (e.g., security, performance, API design, data modeling) and/or a feature name. If neither is clear, ask up to 3 targeted questions:

- What feature is this checklist for? - What quality domain should it focus on? - Are there specific concerns from the constitution or prior analysis?

1. Locate artifacts: Load spec.md for the identified feature, plus plan.md and tasks.md if available. Load .specify/memory/constitution.md.
2. Generate FEATURE_DIR/checklists/<domain>.md: # Quality Checklist: <Domain> — <Feature Name> **Generated**: <YYYY-MM-DD> **Scope**: Requirements quality validation (not implementation verification) ## Completeness - [] CHK001 Are all user scenarios covered by at least one functional requirement? [Completeness, Spec §User Scenarios] - [] CHK002 Does each functional requirement have a corresponding success criterion? [Completeness, Spec §Success Criteria] ## Clarity - [] CHK010 Are all vague terms replaced with measurable criteria? [Clarity, Gap] - [] CHK011 Is every acronym and domain term defined? [Clarity, Spec §Key Entities] ## Consistency - [] CHK020 Is the same term used consistently for the same concept throughout? [Consistency] ## Coverage - [] CHK030 Are error conditions addressed for each user scenario? [Gap] - [] CHK031 Are non-functional requirements specified? [Gap] ## Measurability - [] CHK040 Does each success criterion include a specific metric or verifiable condition? [Measurability, Spec §Success Criteria] Item rules:

- Ask about requirement quality, never implementation success. Bad: "Verify the button works." Good: "Is the expected button behavior described for each user scenario?" - Include a quality dimension in brackets: [Completeness], [Clarity], [Consistency], [Gap], [Ambiguity], [Conflict], [Measurability] - Reference specific spec sections: [Spec §X.Y] - At least 80% of items must trace to a specific spec section or an identified gap

1. Report: File path, total item count, breakdown by quality dimension.

Next step: Complete all checklist items before running /speckit.implement.

{{INPUT}}