Skip to content
Back to skills

Speckit Checklist

ASecurity

Checklist focus (consider it if not empty): $ARGUMENTS A checklist tests whether the **requirements are well written** — complete, clear, consistent, measurable, covering the scenarios — not whether the implementation works. - ✅ "Is 'prominent display' quantified with specific sizing? [Clarity, Spec §FR-4]" - ✅ "Are error responses specified for every failure mode? [Completeness, Gap]" - ❌ "Verify the button works" / "Test error handling" / "Confirm the API returns 200"

  • 3 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added September 25, 2026
ai-agentsbashapisecurityperformance

Works with

  • api

Security analysis

A100/100

Scanned September 25, 2026

npx -y skills add rayedbajwa/spaces --skill speckit-checklist --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Speckit Checklist?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Speckit Checklist
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/rayedbajwa-speckit-checklist/badge)](https://www.skillsdirectory.com/skills/rayedbajwa-speckit-checklist)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
# Checklist: unit tests for the requirements

Checklist focus (consider it if not empty): $ARGUMENTS

A checklist tests whether the **requirements are well written** — complete, clear, consistent, measurable, covering the scenarios — not whether the implementation works.

- ✅ "Is 'prominent display' quantified with specific sizing? [Clarity, Spec §FR-4]"
- ✅ "Are error responses specified for every failure mode? [Completeness, Gap]"
- ❌ "Verify the button works" / "Test error handling" / "Confirm the API returns 200"

## Steps

1. Run `.specify/scripts/bash/check-prerequisites.sh --json`; use `FEATURE_DIR` and `AVAILABLE_DOCS`.
2. **Scope**: from the focus above and the spec/plan/tasks, decide the domain (e.g. api, ux, security, performance; the focus above names it when given), depth (default: standard) and audience (default: PR reviewer). Ask only when the answer changes the checklist materially — at most 3 questions, as `## Question N: <topic>` with a short options table (Option | Candidate | Why it matters), then stop and wait. Otherwise use the defaults and the top two focus areas.
3. Read the parts of `spec.md` (and `plan.md`, `tasks.md` when they exist) that matter for the focus.
4. Write `FEATURE_DIR/checklists/<domain>.md` from `.specify/templates/checklist-template.md`. If the file exists, append and continue its CHK numbering; never delete items.
   - Group by: Requirement Completeness · Clarity · Consistency · Acceptance Criteria Quality · Scenario Coverage (primary, alternate, exception, recovery) · Edge Cases · Non-Functional · Dependencies & Assumptions · Ambiguities & Conflicts.
   - Each item: `- [ ] CHK001 <question about the requirement> [<dimension>, Spec §X]`. At least 80% cite a spec section or a marker: `[Gap]`, `[Ambiguity]`, `[Conflict]`, `[Assumption]`.
   - Phrase items as "Are … defined/specified for …?", "Is … quantified?", "Are … consistent between … and …?", "Can … be objectively measured?". Never start with Verify/Test/Confirm/Check an implementation behaviour, and no implementation details.
   - Rollback/recovery requirements when state changes. Merge near-duplicates; at most ~40 items, highest risk first; fold minor edge cases into one item.
5. **Report**: checklist path, item count, created or appended, focus areas, depth and audience.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.