speckit-critique-run
Skill548 starsChanged 12 days agoApache-2.0View on GitHub
Perform a dual-lens critical review of the specification and plan from
In this repository54 other files it gives its agents.
- AGENTS.mdAGENTS.md
- backend/AGENTS.mdAGENTS.md
- docs/AGENTS.mdAGENTS.md
- backend/CLAUDE.mdCLAUDE.md
- CLAUDE.mdCLAUDE.md
- docs/CLAUDE.mdCLAUDE.md
- analyzing-ci-flakinessSkill
- audit-docsSkill
- commitSkill
- creating-changelog-entriesSkill
- creating-issuesSkill
- creating-prdSkill
- grilling-ideasSkill
- harvesting-reviewSkill
- migrate-feature-pageSkill
- monitoring-pull-requestsSkill
- neo4j-cypher-guideSkill
- opsmill-dev-analyzing-bugsSkill
- opsmill-dev-fixing-bugsSkill
- opsmill-dev-test-driving-bugsSkill
- pre-ciSkill
- prSkill
- rebaseSkill
- reviewing-pull-requestsSkill
- speckit-agent-context-updateSkill
- speckit-analyzeSkill
- speckit-checklistSkill
- speckit-clarifySkill
- speckit-constitutionSkill
- speckit-convergeSkill
- speckit-git-commitSkill
- speckit-git-featureSkill
- speckit-git-initializeSkill
- speckit-git-remoteSkill
- speckit-git-validateSkill
- speckit-implementSkill
- speckit-opsmill-autoSkill
- speckit-opsmill-extractSkill
- speckit-opsmill-implementSkill
- speckit-opsmill-prepSkill
- speckit-opsmill-retrospectSkill
- speckit-opsmill-summarySkill
- speckit-planSkill
- speckit-review-codeSkill
- speckit-review-commentsSkill
- speckit-review-errorsSkill
- speckit-review-runSkill
- speckit-review-simplifySkill
- speckit-review-testsSkill
- speckit-review-typesSkill
- speckit-specifySkill
- speckit-tasksSkill
- speckit-taskstoissuesSkill
- upgrading-infrastructure-dependenciesSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
20 rules. About 0 are common, 4 are distinctive, and 16 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
16
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 22
Check if .specify/extensions.yml exists in the project root.
- Very specific: line 23
If it exists, read it and look for entries under the hooks.before_critique key
- Distinctive: line 24
If the YAML cannot be parsed or is invalid, skip hook checking silently and continue normally
- Very specific: line 27
If the hook has no condition field, or it is null/empty, treat the hook as executable
- Very specific: line 28
If the hook defines a non-empty condition, skip the hook and leave condition evaluation to the HookExecutor implementation
- Very specific: line 51
If no hooks are registered or .specify/extensions.yml does not exist, skip silently
- Very specific: line 67
Run .specify/scripts/bash/check-prerequisites.sh --json --include-tasks from repo root and parse FEATURE_DIR and AVAILABLE_DOCS list.
- Very specific: line 70
REQUIRED: Read spec.md for requirements, user stories, and acceptance criteria
- Very specific: line 71
REQUIRED: Read plan.md for architecture, tech stack, and implementation phases
- Very specific: line 72
IF EXISTS: Read .specify/memory/constitution.md for governing principles
- Very specific: line 73
IF EXISTS: Read tasks.md for task breakdown (if already generated)
- Very specific: line 74
IF EXISTS: Read previous critique reports in FEATURE_DIR/critiques/ for context
- Very specific: line 178
Ensure the directory FEATURE_DIR/critiques/ exists (create it if necessary), then create the critique report at …
- Distinctive: line 184
Findings Summary Table: All items with ID, lens, severity, summary, suggestion
- Distinctive: line 202
Ask: "Would you like me to apply these changes?
- Distinctive: line 203
If user approves, apply changes to the relevant files
- Very specific: line 204
After applying changes, recommend re-running /speckit.critique to verify
- Very specific: line 209
If PROCEED: "Run /speckit.tasks to break the plan into actionable tasks"
- Very specific: line 210
If PROCEED WITH UPDATES: "Review the suggested changes, then run /speckit.tasks"
- Very specific: line 211
If RETHINK: "Consider running /speckit.specify to refine the spec or /speckit.plan to revise the architecture"
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
User Input
$ARGUMENTSYou MUST consider the user input before proceeding (if not empty).
Pre-Execution Checks
Check for extension hooks (before critique):
- Check if
.specify/extensions.ymlexists in the project root. - If it exists, read it and look for entries under the
hooks.before_critiquekey - If the YAML cannot be parsed or is invalid, skip hook checking silently and continue normally
- Filter out hooks where
enabledis explicitlyfalse. Treat hooks without anenabledfield as enabled by default. - For each remaining hook, do not attempt to interpret or evaluate hook
conditionexpressions:- If the hook has no
conditionfield, or it is null/empty, treat the hook as executable - If the hook defines a non-empty
condition, skip the hook and leave condition evaluation to the HookExecutor implementation
- If the hook has no
- For each executable hook, output the following based on its
optionalflag:- Optional hook (
optional: true):## Extension Hooks **Optional Pre-Hook**: {extension} Command: `/{command}` Description: {description} Prompt: {prompt} To execute: `/{command}` - Mandatory hook (
optional: false):## Extension Hooks **Automatic Pre-Hook**: {extension} Executing: `/{command}` EXECUTE_COMMAND: {command} Wait for the result of the hook command before proceeding to the Outline.
- Optional hook (
- If no hooks are registered or
.specify/extensions.ymldoes not exist, skip silently
Goal
Challenge the specification and implementation plan through two distinct expert lenses BEFORE committing to implementation. The Product Lens evaluates whether the right problem is being solved in the right way for users. The Engineering Lens evaluates whether the technical approach is sound, scalable, and free of hidden risks. This dual review prevents costly mid-implementation pivots and catches strategic and technical blind spots early.
Operating Constraints
STRICTLY READ-ONLY FOR EXISTING ARTIFACTS: During this command, do not directly modify existing project files such as spec.md, plan.md, or other source/docs. You may create a new critique report under FEATURE_DIR/critiques/critique-{timestamp}.md. Propose, but do not apply, edits to spec.md/plan.md; applying any changes requires explicit user approval in a follow-up step or command after the user reviews the findings.
CONSTRUCTIVE CHALLENGE: The goal is to strengthen the spec and plan, not to block progress. Every critique item must include a constructive suggestion for improvement.
Constitution Authority: The project constitution (.specify/memory/constitution.md) defines non-negotiable principles. Any spec/plan element conflicting with the constitution is automatically a 🎯 Must-Address item.
Outline
- Run
.specify/scripts/bash/check-prerequisites.sh --json --include-tasksfrom repo root and parse FEATURE_DIR and AVAILABLE_DOCS list. All paths must be absolute. For single quotes in args like "I'm Groot", use escape syntax: e.g 'I'''m Groot' (or double-quote if possible: "I'm Groot"). - Load Critique Context:
- REQUIRED: Read
spec.mdfor requirements, user stories, and acceptance criteria - REQUIRED: Read
plan.mdfor architecture, tech stack, and implementation phases - IF EXISTS: Read
.specify/memory/constitution.mdfor governing principles - IF EXISTS: Read
tasks.mdfor task breakdown (if already generated) - IF EXISTS: Read previous critique reports in FEATURE_DIR/critiques/ for context
- REQUIRED: Read
Product Lens Review (CEO/Product Lead Perspective):
Adopt the mindset of an experienced product leader who cares deeply about user value, market fit, and business impact. Evaluate:
3a. Problem Validation
- Is the problem statement clear and well-defined?
- Is this solving a real user pain point, or is it a solution looking for a problem?
- What evidence supports the need for this feature? (user research, data, customer requests)
- Is the scope appropriate — not too broad (trying to do everything) or too narrow (missing the core value)?
3b. User Value Assessment
- Does every user story deliver tangible user value?
- Are the acceptance criteria written from the user's perspective (outcomes, not implementation)?
- Is the user journey complete — or are there gaps where users would get stuck?
- What's the simplest version that would deliver 80% of the value? (MVP analysis)
- Are there unnecessary features that add complexity without proportional value?
3c. Alternative Approaches
- Could a simpler solution achieve the same outcome?
- Are there existing tools, libraries, or services that could replace custom implementation?
- What would a competitor's approach look like?
- What would happen if this feature were NOT built? What's the cost of inaction?
3d. Edge Cases & User Experience
- What happens when things go wrong? (error states, empty states, loading states)
- How does this feature interact with existing functionality?
- Are accessibility considerations addressed?
- Is the feature discoverable and intuitive?
- What are the onboarding/migration implications for existing users?
3e. Success Measurement
- Are the success criteria measurable and time-bound?
- How will you know if this feature is successful after launch?
- What metrics should be tracked?
- What would trigger a rollback decision?
Engineering Lens Review (Staff Engineer Perspective):
Adopt the mindset of a senior staff engineer who has seen projects fail due to hidden technical risks. Evaluate:
4a. Architecture Soundness
- Does the architecture follow established patterns for this type of system?
- Are boundaries and interfaces well-defined (separation of concerns)?
- Is the architecture testable at each layer?
- Are there circular dependencies or tight coupling risks?
- Does the architecture support future evolution without major refactoring?
4b. Failure Mode Analysis
- What are the most likely failure modes? (network failures, data corruption, resource exhaustion)
- How does the system degrade gracefully under each failure mode?
- What happens under peak load? Is there a scaling bottleneck?
- What are the blast radius implications — can a failure in this feature affect other parts of the system?
- Are retry, timeout, and circuit-breaker strategies defined?
4c. Security & Privacy Review
- What is the threat model? What attack vectors does this feature introduce?
- Are trust boundaries clearly defined (user input, API responses, third-party data)?
- Is sensitive data handled appropriately (encryption, access control, retention)?
- Are there compliance implications (GDPR, SOC2, HIPAA)?
- Is the principle of least privilege followed?
4d. Performance & Scalability
- Are there potential bottlenecks in the data flow?
- What are the expected data volumes? Will the design handle 10x growth?
- Are caching strategies appropriate and cache invalidation well-defined?
- Are database queries optimized (indexing, pagination, query complexity)?
- Are there resource-intensive operations that should be async or batched?
4e. Testing Strategy
- Is the testing plan comprehensive (unit, integration, E2E)?
- Are the critical paths identified for priority testing?
- Is the test data strategy realistic?
- Are there testability concerns (hard-to-mock dependencies, race conditions)?
- Is the test coverage target appropriate for the risk level?
4f. Operational Readiness
- Is observability planned (logging, metrics, tracing)?
- Are alerting thresholds defined?
- Is there a rollback strategy?
- Are database migrations reversible?
- Is the deployment strategy clear (blue-green, canary, feature flags)?
4g. Dependencies & Integration Risks
- Are third-party dependencies well-understood (stability, licensing, maintenance)?
- Are integration points with existing systems well-defined?
- What happens if an external service is unavailable?
- Are API versioning and backward compatibility considered?
- Cross-Lens Synthesis:
Identify items where both lenses converge (these are highest priority):
- Product simplification that also reduces engineering risk
- Engineering constraints that affect user experience
- Scope adjustments that improve both value delivery and technical feasibility
Severity Classification: Classify each finding:
- 🎯 Must-Address: Blocks proceeding to implementation. Critical product gap, security vulnerability, architecture flaw, or constitution violation. Must be resolved before
/speckit.tasks. - 💡 Recommendation: Strongly suggested improvement that would significantly improve quality, value, or risk profile. Should be addressed but won't block progress.
- 🤔 Question: Ambiguity or assumption that needs stakeholder input. Cannot be resolved by the development team alone.
- 🎯 Must-Address: Blocks proceeding to implementation. Critical product gap, security vulnerability, architecture flaw, or constitution violation. Must be resolved before
Generate Critique Report: Ensure the directory
FEATURE_DIR/critiques/exists (create it if necessary), then create the critique report atFEATURE_DIR/critiques/critique-{timestamp}.mdusing.specify/templates/critique-template.mdas the required structure. The report must include:- Executive Summary: Overall assessment and readiness to proceed
- Product Lens Findings: Organized by subcategory (3a-3e)
- Engineering Lens Findings: Organized by subcategory (4a-4g)
- Cross-Lens Insights: Items where both perspectives converge
- Findings Summary Table: All items with ID, lens, severity, summary, suggestion
Findings Table Format:
ID Lens Severity Category Finding Suggestion P1 Product 🎯 Problem Validation No evidence of user need Conduct 5 user interviews or reference support tickets E1 Engineering 💡 Failure Modes No retry strategy for API calls Add exponential backoff with circuit breaker X1 Both 🎯 Scope × Risk Feature X adds complexity with unclear value Defer to v2; reduces both scope and technical risk - Provide Verdict:
Based on findings, provide one of:
- ✅ PROCEED: No must-address items. Spec and plan are solid. Run
/speckit.tasksto proceed. - ⚠️ PROCEED WITH UPDATES: Must-address items found but are resolvable. Offer to apply fixes to spec/plan, then proceed.
- 🛑 RETHINK: Fundamental product or architecture concerns. Recommend revisiting the spec with
/speckit.specifyor the plan with/speckit.plan.
- ✅ PROCEED: No must-address items. Spec and plan are solid. Run
- Offer Remediation:
For each must-address item and recommendation:
- Provide a specific suggested edit to
spec.mdorplan.md - Ask: "Would you like me to apply these changes? (all / select / none)"
- If user approves, apply changes to the relevant files
- After applying changes, recommend re-running
/speckit.critiqueto verify
- Provide a specific suggested edit to
Post-Critique Actions
Suggest next steps based on verdict:
- If PROCEED: "Run
/speckit.tasksto break the plan into actionable tasks" - If PROCEED WITH UPDATES: "Review the suggested changes, then run
/speckit.tasks" - If RETHINK: "Consider running
/speckit.specifyto refine the spec or/speckit.planto revise the architecture"
Check for extension hooks (after critique):
- Check if
.specify/extensions.ymlexists in the project root. - If it exists, read it and look for entries under the
hooks.after_critiquekey - If the YAML cannot be parsed or is invalid, skip hook checking silently and continue normally
- Filter out hooks where
enabledis explicitlyfalse. Treat hooks without anenabledfield as enabled by default. - For each remaining hook, do not attempt to interpret or evaluate hook
conditionexpressions:- If the hook has no
conditionfield, or it is null/empty, treat the hook as executable - If the hook defines a non-empty
condition, skip the hook and leave condition evaluation to the HookExecutor implementation
- If the hook has no
- For each executable hook, output the following based on its
optionalflag:- Optional hook (
optional: true):## Extension Hooks **Optional Hook**: {extension} Command: `/{command}` Description: {description} Prompt: {prompt} To execute: `/{command}` - Mandatory hook (
optional: false):## Extension Hooks **Automatic Hook**: {extension} Executing: `/{command}` EXECUTE_COMMAND: {command}
- Optional hook (
- If no hooks are registered or
.specify/extensions.ymldoes not exist, skip silently
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

