skills/spec-review/SKILL.md
Review a spec document for completeness before approval. Use after interview, before implementation, or when asked to 'review spec', 'check spec', or 'is this spec ready'. Read-only analysis that flags gaps.
npx skillsauth add opsmachine/om-agency spec-reviewInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Reviews a specification document for completeness and quality before approval. This is a read-only skill that analyzes without modifying.
Use this skill when:
/spec-review path/to/spec.mdRead the spec document at the path provided by the user.
See shared/spec-io.md for spec file location and section structure.
If no path provided, look for recent specs in Documents/specs/ and ask which one to review.
Gate check: Verify **Status:** is Draft. If already Approved, the spec has been reviewed — nothing to do here.
Evaluate each section against these criteria:
Red flags:
Red flags:
Red flags:
Red flags:
Red flags:
Note: This section helps activate falsification (testing if you're wrong BEFORE implementing). It prevents discovering issues through iteration that could be caught upfront.
Acceptable if empty - not all specs need technical notes
Red flags:
See
shared/security-lens.mdfor what good answers look like.
Acceptable if empty - not all interviews surface questions
Rate the spec on this scale:
Ready for Implementation
Needs Minor Revision
Needs Major Revision
Format your response as:
## Spec Review: {spec name}
### Readiness: {Ready | Needs Minor Revision | Needs Major Revision}
### Completeness Checklist
**Problem Statement:** {Complete | Incomplete}
- {specific feedback}
**Acceptance Criteria:** {Complete | Incomplete}
- {specific feedback}
- {count} criteria found
**Non-Goals:** {Complete | Incomplete}
- {specific feedback}
**Assumptions:** {Complete | Incomplete}
- {specific feedback}
**Falsification Analysis:** {Complete | Incomplete}
- {specific feedback}
- Failure modes considered: {yes/no}
**Security Considerations:** {Complete | Incomplete}
- {specific feedback}
### Gaps Found
{List specific gaps if any}
### Suggested Questions
{Questions that might have been missed during interview}
### Recommendation
{Clear next step - approve and proceed, or revise specific sections}
This skill is read-only. Do not:
If revisions are needed, tell the user what to change and let them (or the interview skill) make the updates.
User: /spec-review Documents/specs/42-dark-mode-spec.md
Claude:
## Spec Review: Dark Mode Toggle
### Readiness: Needs Minor Revision
### Completeness Checklist
**Problem Statement:** Complete
- Clear problem: users want dark mode for eye strain
- Impact: affects all users, especially night-time use
- Beneficiary: all application users
**Acceptance Criteria:** Complete
- 4 testable criteria found
- Edge cases covered (persistence, browser support)
**Non-Goals:** Incomplete
- Only says "not doing system-wide theme"
- Should clarify: custom color themes? Per-page themes?
**Assumptions:** Complete
- Assumes CSS variables already in place
- Assumes local storage is acceptable for persistence
### Gaps Found
1. Non-Goals could be more specific about what theme features are deferred
### Suggested Questions
- Should dark mode sync across devices (if user is logged in)?
- What's the default for new users - light or dark?
### Recommendation
Minor revision needed. Add specifics to Non-Goals section and consider the syncing question. Otherwise ready to proceed.
testing
Write failing tests for all planned acceptance criteria from the test plan. Use after /plan-tests, before implementation. Invoke with '/write-failing-test path/to/spec.md' or 'write failing test', 'red phase', 'start TDD'.
data-ai
Workflow manager that orchestrates the entire skill system. Runs automatically before any implementation work. Reads state from artifacts, determines the next skill, spawns sub-agents for execution, and manages human gates. Invoke with '/workflow-router' or it runs automatically per CLAUDE.md.
tools
Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.
development
Supabase security best practices and patterns. Use when working with Supabase projects, creating tables, writing RLS policies, edge functions, or reviewing Supabase code. Invoke with '/supabase-security' or when asked about Supabase security.