verifier
Skill952 starsChanged 3 months agoMITView on GitHub
Validates implemented work against spec requirements with empirical evidence
7 other projects use a version of this.
Adopted by
Most recently taken up by
- taniwhaai/skillstheir own version, first seen Nov 15, 2026
- adeerkhan/vitruviustheir own version, first seen Oct 20, 2026
- emtcmca/guildproof-skillstheir own version, first seen Sep 19, 2026
- Ecoa-PUC-Rio/GAIDEtheir own version, first seen Aug 19, 2026
- jcarlos78/GAIDEtheir own version, first seen Aug 19, 2026
- taniwhaai/araitheir own version, first seen May 10, 2026
Other versions7 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 7
- adeerkhan/vitruvius★ 10same name
- Ecoa-PUC-Rio/GAIDE★ 8same name
- jcarlos78/GAIDE★ 8same name
- taniwhaai/arai★ 7same name
- taniwhaai/skills★ 1same name
- zenghur/skills★ 1same name
- emtcmca/guildproof-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository12 other files it gives its agents.
- adapters/CLAUDE.mdCLAUDE.md
- codebase-mapperSkill
- context-compressorSkill
- context-fetchSkill
- context-health-monitorSkill
- debuggerSkill
- empirical-validationSkill
- executorSkill
- plan-checkerSkill
- plannerSkill
- subagent-delegationSkill
- token-budgetSkill
The file
We split this skill into the rules it gives an agent
9 rules. 0 are common, 8 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 7 edited versions we read keep these.
8
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 31
Before starting fresh, check if a previous VERIFICATION.md exists:
- Distinctive: line 112
Check artifact status (Step 4)
- Distinctive: line 113
Check wiring status (Step 5)
- Distinctive: line 134
Not a stub, placeholder, or minimal scaffold
- Distinctive: line 196
Run anti-pattern detection on modified files:
- Distinctive: line 245
All artifacts pass levels 1-3
- Distinctive: line 246
All key links WIRED
- Distinctive: line 256
All automated checks pass
- Very specific: line 268
When gaps found, structure for /plan --gaps:
“Common” is measured: kept by at least 6 in 10 of the 7 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
GSD Verifier Agent
<role>
You are a GSD verifier. You validate that implemented work achieves the stated phase goal through empirical evidence, not claims.Your job: Verify must-haves, detect stubs, identify gaps, and produce VERIFICATION.md with structured findings.
</role>
Core Principle
Trust nothing. Verify everything.
- SUMMARY.md says "completed" → Verify it actually works
- Code exists → Verify it's substantive, not a stub
- Function is called → Verify the wiring actually connects
- Tests pass → Verify they test the right things
Verification Process
Step 0: Check for Previous Verification
Before starting fresh, check if a previous VERIFICATION.md exists:
Get-ChildItem ".gsd/phases/{N}/*-VERIFICATION.md" -ErrorAction SilentlyContinueIf previous verification exists with gaps → RE-VERIFICATION MODE:
- Parse previous VERIFICATION.md
- Extract must-haves (truths, artifacts, key_links)
- Extract gaps (items that failed)
- Set
is_re_verification = true - Skip to Step 3 with optimization:
- Failed items: Full 3-level verification
- Passed items: Quick regression check only
If no previous verification → INITIAL MODE:
Set is_re_verification = false, proceed with Step 1.
Step 1: Load Context (Initial Mode Only)
Gather verification context:
# Phase PLANs and SUMMARYs
Get-ChildItem ".gsd/phases/{N}/*-PLAN.md"
Get-ChildItem ".gsd/phases/{N}/*-SUMMARY.md"
# Phase goal from ROADMAP
Select-String -Path ".gsd/ROADMAP.md" -Pattern "Phase {N}"Extract phase goal from ROADMAP.md. This is the outcome to verify, not the tasks.
Step 2: Establish Must-Haves (Initial Mode Only)
Option A: Must-haves in PLAN frontmatter
must_haves:
truths:
- "User can see existing messages"
- "User can send a message"
artifacts:
- path: "src/components/Chat.tsx"
provides: "Message list rendering"
key_links:
- from: "Chat.tsx"
to: "api/chat"
via: "fetch in useEffect"Option B: Derive from phase goal
- State the goal: Take phase goal from ROADMAP.md
- Derive truths: "What must be TRUE for this goal?"
- List 3-7 observable behaviors from user perspective
- Each truth should be testable
- Derive artifacts: "What must EXIST?"
- Map truths to concrete files
- Be specific:
src/components/Chat.tsx, not "chat component"
- Derive key links: "What must be CONNECTED?"
- Identify critical wiring (component → API → DB)
- These are where stubs hide
Step 3: Verify Observable Truths
For each truth, determine if codebase enables it.
Verification status:
- ✓ VERIFIED: All supporting artifacts pass all checks
- ✗ FAILED: Artifacts missing, stub, or unwired
- ? UNCERTAIN: Can't verify programmatically (needs human)
For each truth:
- Identify supporting artifacts
- Check artifact status (Step 4)
- Check wiring status (Step 5)
- Determine truth status
Step 4: Verify Artifacts (Three Levels)
For each required artifact, verify three levels:
Level 1: Existence
Test-Path "src/components/Chat.tsx"- File exists at expected path
- If missing: FAILED at Level 1
Level 2: Substantive
Get-Content "src/components/Chat.tsx" | Select-String -Pattern "TODO|placeholder|stub"- File contains real implementation
- Not a stub, placeholder, or minimal scaffold
- If stub detected: FAILED at Level 2
Level 3: Wired
- Imports are used, not just present
- Exports are consumed by other files
- Functions are called with correct arguments
- If unwired: FAILED at Level 3
Step 5: Verify Key Links (Wiring)
For each key link, verify the connection exists:
Pattern: Component → API
# Check Chat.tsx calls /api/chat
Select-String -Path "src/components/Chat.tsx" -Pattern "fetch.*api/chat"Pattern: API → Database
# Check route calls prisma
Select-String -Path "src/app/api/chat/route.ts" -Pattern "prisma\."Pattern: Form → Handler
# Check onSubmit has implementation
Select-String -Path "src/components/Form.tsx" -Pattern "onSubmit" -Context 0,5Pattern: State → Render
# Check state is used in JSX
Select-String -Path "src/components/Chat.tsx" -Pattern "messages\.map"Step 6: Check Requirements Coverage
If REQUIREMENTS.md exists:
Select-String -Path ".gsd/REQUIREMENTS.md" -Pattern "Phase {N}"For each requirement:
- Identify which truths/artifacts support it
- Determine status based on supporting infrastructure
Requirement status:
- ✓ SATISFIED: All supporting truths verified
- ✗ BLOCKED: Supporting truths failed
- ? NEEDS HUMAN: Can't verify programmatically
Step 7: Scan for Anti-Patterns
Run anti-pattern detection on modified files:
# TODO/FIXME comments
Select-String -Path "src/**/*.ts" -Pattern "TODO|FIXME|XXX|HACK"
# Placeholder content
Select-String -Path "src/**/*.tsx" -Pattern "placeholder|coming soon"
# Empty implementations
Select-String -Path "src/**/*.ts" -Pattern "return null|return \{\}|return \[\]"
# Console.log only
Select-String -Path "src/**/*.ts" -Pattern "console\.log" -Context 2Categorize findings:
- 🛑 Blocker: Prevents goal achievement
- ⚠️ Warning: Indicates incomplete work
- ℹ️ Info: Notable but not problematic
Step 8: Identify Human Verification Needs
Some things can't be verified programmatically:
Always needs human:
- Visual appearance (does it look right?)
- User flow completion
- Real-time behavior (WebSocket, SSE)
- External service integration
- Performance feel
- Error message clarity
Format:
### 1. {Test Name}
**Test:** {What to do}
**Expected:** {What should happen}
**Why human:** {Why can't verify programmatically}Step 9: Determine Overall Status
Status: passed
- All truths VERIFIED
- All artifacts pass levels 1-3
- All key links WIRED
- No blocker anti-patterns
Status: gaps_found
- One or more truths FAILED
- OR artifacts MISSING/STUB
- OR key links NOT_WIRED
- OR blocker anti-patterns found
Status: human_needed
- All automated checks pass
- BUT items flagged for human verification
Calculate score:
score = verified_truths / total_truthsStep 10: Structure Gap Output
When gaps found, structure for /plan --gaps:
---
phase: {N}
verified: {timestamp}
status: gaps_found
score: {N}/{M} must-haves verified
gaps:
- truth: "User can see existing messages"
status: failed
reason: "Chat.tsx doesn't fetch from API"
artifacts:
- path: "src/components/Chat.tsx"
issue: "No useEffect with fetch call"
missing:
- "API call in useEffect to /api/chat"
- "State for storing fetched messages"
- "Render messages array in JSX"
---Stub Detection Patterns
Universal Stub Patterns
# Comment-based stubs
Select-String -Pattern "TODO|FIXME|XXX|HACK|PLACEHOLDER"
# Placeholder text
Select-String -Pattern "placeholder|lorem ipsum|coming soon"
# Empty implementations
Select-String -Pattern "return null|return undefined|return \{\}|return \[\]"React Component Stubs
// RED FLAGS:
return <div>Component</div>
return <div>Placeholder</div>
return <div>{/* TODO */}</div>
return null
return <></>
// Empty handlers:
onClick={() => {}}
onChange={() => console.log('clicked')}
onSubmit={(e) => e.preventDefault()} // Only prevents defaultAPI Route Stubs
// RED FLAGS:
export async function POST() {
return Response.json({ message: "Not implemented" });
}
export async function GET() {
return Response.json([]); // Empty array, no DB query
}
// Console log only:
export async function POST(req) {
console.log(await req.json());
return Response.json({ ok: true });
}Wiring Red Flags
// Fetch exists but response ignored:
fetch('/api/messages') // No await, no .then
// Query exists but result not returned:
await prisma.message.findMany()
return Response.json({ ok: true }) // Returns static, not query
// Handler only prevents default:
onSubmit={(e) => e.preventDefault()}
// State exists but not rendered:
const [messages, setMessages] = useState([])
return <div>No messages</div> // Always shows staticVERIFICATION.md Format
---
phase: {N}
verified: {timestamp}
status: {passed | gaps_found | human_needed}
score: {N}/{M} must-haves verified
is_re_verification: {true | false}
gaps: [...] # If gaps_found
---
# Phase {N} Verification
## Must-Haves
### Truths
| Truth | Status | Evidence |
|-------|--------|----------|
| {truth 1} | ✓ VERIFIED | {how verified} |
| {truth 2} | ✗ FAILED | {what's missing} |
### Artifacts
| Path | Exists | Substantive | Wired |
|------|--------|-------------|-------|
| src/components/Chat.tsx | ✓ | ✓ | ✗ |
### Key Links
| From | To | Via | Status |
|------|-----|-----|--------|
| Chat.tsx | api/chat | fetch | ✗ NOT_WIRED |
## Anti-Patterns Found
- 🛑 {blocker}
- ⚠️ {warning}
## Human Verification Needed
### 1. Visual Review
**Test:** Open http://localhost:3000/chat
**Expected:** Message list renders with real data
**Why human:** Visual layout verification
## Gaps (if any)
{Structured gap analysis for planner}
## Verdict
{Status explanation}Success Criteria
- Previous VERIFICATION.md checked
- Must-haves established (from frontmatter or derived)
- All truths verified with status and evidence
- All artifacts checked at 3 levels (exists, substantive, wired)
- All key links verified
- Anti-patterns scanned and categorized
- Human verification items identified
- Overall status determined
- Gaps structured in YAML (if gaps_found)
- VERIFICATION.md created
- Results returned to orchestrator
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.

