Verification Loop

Verification Loop

A comprehensive verification system for ensuring code quality across multiple dimensions.

When to Use

Invoke this skill:

• At task start — to declare the verification surface before writing code (Phase 0)
• After completing a feature or significant code change
• Before creating a pull request
• When you want to ensure quality gates pass
• After refactoring or major modifications
• During code review preparation

Verification Phases

Phase 0: Declare the verification surface (BEFORE any work)

Run this before writing code, not after. The other phases ratify finished work; this one decides — up front — how that work will be proven, and ensures the proof comes from outside the model.

State three things before starting:

1. The external signal. Name the concrete command or surface that will confirm success: a Bash test/lint/typecheck command, an MCP browser loading the real page and reading the DOM/console, a fresh session, or a second agent. "I'll review it myself when done" is not a verification surface.
2. Why it is external to the generating context. A self-check inside the same context that produced the code can ratify the very error that context introduced. The signal must come from outside that context — a process the model cannot silently satisfy by asserting success. (Origin failure mode: a model asked to fix a wrong display edited the underlying value and marked the task resolved — the system failed and reported success. Every phase below exists to make that signature visible; this phase makes it visible before it spreads.)
3. If no surface exists, building it is part of the task. Do not proceed to implementation until the signal is wired. The verification surface is a deliverable, not an afterthought.

Declaration stamp: Verify-by: <command/surface> — external? Y/N — exists yet? Y/N. If exists yet = N, the first build step is to create it.

This is the forward-declaration that Phases 1–6 assume. It is the FRAME-phase half of this skill; Phases 1–6 are the PROVE-phase half. (Constitutional anchor: Reliquary Rule #64(c) "govern the model, don't execute through it"; operator-invoked equivalent: /v-govern step 3.)

Phase 1: Build Verification

# Check if project builds
npm run build 2>&1 | tail -20
# OR
pnpm build 2>&1 | tail -20
# OR for other build systems
make build 2>&1 | tail -20
cargo build 2>&1 | tail -20

Critical: If build fails, STOP and fix before continuing.

Phase 2: Type Check

# TypeScript projects
npx tsc --noEmit 2>&1 | head -30

# Python projects
pyright . 2>&1 | head -30
mypy . 2>&1 | head -30

# Go projects
go vet ./... 2>&1 | head -30

Report all type errors. Fix critical ones before continuing.

Phase 3: Lint Check

# JavaScript/TypeScript
npm run lint 2>&1 | head -30
eslint . 2>&1 | head -30

# Python
ruff check . 2>&1 | head-30
flake8 . 2>&1 | head -30

# Go
golint ./... 2>&1 | head -30

# Rust
cargo clippy 2>&1 | head -30

Phase 4: Test Suite

# Run tests with coverage
npm run test -- --coverage 2>&1 | tail -50
pytest --cov=. 2>&1 | tail -50
go test -cover ./... 2>&1 | tail -50

# Check coverage threshold
# Target: 80% minimum for new code

Report:

• Total tests: X
• Passed: X
• Failed: X
• Coverage: X%
• New files covered: X%

Phase 5: Security Scan

# Check for secrets
grep -rn "sk-" --include="*.ts" --include="*.js" --include="*.py" . 2>/dev/null | head -10
grep -rn "api_key" --include="*.ts" --include="*.js" --include="*.py" . 2>/dev/null | head -10
grep -rn "password.*=" --include="*.ts" --include="*.js" --include="*.py" . 2>/dev/null | head -10

# Check for debug statements
grep -rn "console.log" --include="*.ts" --include="*.tsx" --include="*.js" src/ 2>/dev/null | head -10
grep -rn "print(" --include="*.py" . 2>/dev/null | head -10
grep -rn "println!" --include="*.rs" src/ 2>/dev/null | head -10

Phase 6: Diff Review

# Show what changed
git diff --stat
git diff HEAD~1 --name-only
git diff --shortstat

Review each changed file for:

• Unintended changes or modifications
• Missing error handling
• Potential edge cases
• Documentation updates needed

Output Format

After running all phases, produce a verification report:

VERIFICATION REPORT
==================

Build:     [✓ PASS / ✗ FAIL]
Types:     [✓ PASS / ✗ FAIL] (X errors)
Lint:      [✓ PASS / ✗ FAIL] (X warnings)
Tests:     [✓ PASS / ✗ FAIL] (X/Y passed, Z% coverage)
Security:  [✓ PASS / ✗ FAIL] (X issues)
Diff:      [X files changed, +Y lines, -Z lines]

Overall:   [✓ READY / ✗ NOT READY] for PR

Issues to Fix:
1. [Priority] Description
2. [Optional] Description
...

Recommendations:
- Suggestion 1
- Suggestion 2

Continuous Mode

For long sessions, run verification every 15 minutes or after major changes:

Checkpoints:

• After completing each function
• After finishing a component
• Before moving to next task
• After resolving merge conflicts

Integration Points

This skill complements:

• testing-patterns: For test design
• deployment-cicd: For CI pipeline preparation
• github-repository-standards: For PR preparation
• security-threat-modeler: For security review

Quick Verification

For rapid checks during development:

# Minimal verification (30 seconds)
npm run build && npm run lint && npm test

Use full verification before commits.

Related Skills

Complementary Skills (Use Together)

• testing-patterns - Comprehensive testing approaches for the test verification phase
• tdd-workflow - Test-driven development that produces code ready for verification
• deployment-cicd - CI pipelines that automate the verification loop

Alternative Skills (Similar Purpose)

• None - verification-loop is a quality gate process, not an alternative to other practices

Prerequisite Skills (Learn First)

• testing-patterns - Understanding testing helps interpret verification results