Discover
Finds every function, method, type, and file in your repository using language-aware adapters.
Certify
Continuously evaluate every code unit in your repository. Score against versioned policies. Assign time-bound certification you can actually trust.
Why Certify?
Section titled “Why Certify?”CI tells you whether code passes right now.
Certify tells you whether code should still be trusted.
Most engineering teams assume code that once passed review can be trusted indefinitely. That assumption is wrong. Standards evolve, dependencies change, systems grow more complex.
Certify introduces continuous code certification — measurable quality scores with certification that expires intentionally. When certification lapses, code is re-evaluated against current standards.
How It Works
Section titled “How It Works”Evaluate
Collects deterministic evidence from linters, test runners, git history, AST analysis, and code metrics. Scores across 9 quality dimensions.
Certify
Assigns time-bound certification status — certified, probationary, or decertified — with expiration dates that enforce recertification.
Report
Generates a complete report card with per-unit scoring, dimension breakdowns, browsable static site, and a live README badge.
Core Commands
Section titled “Core Commands”certify init # bootstrap config, policies, CI workflowscertify scan # discover all certifiable code unitscertify certify # evaluate, score, and certifycertify report # generate your report cardcertify architect # AI-powered architectural reviewcertify doctor # diagnose setup and configuration issuescertify onboard # interactive step-by-step guide9 Quality Dimensions
Section titled “9 Quality Dimensions”Every code unit is scored across:
| Dimension | What it measures |
|---|---|
| Correctness | Lint errors, vet issues, test failures |
| Maintainability | Cyclomatic complexity, function length |
| Readability | Line length, documentation, TODO count |
| Testability | Test coverage, test existence |
| Security | Security-sensitive patterns (panic, os.Exit) |
| Architectural Fitness | Package structure, dependency patterns |
| Operational Quality | Git churn, contributor count |
| Performance | Algorithmic complexity, nesting depth, quadratic patterns |
| Change Risk | Recent changes, author concentration |
Certification Lifecycle
Section titled “Certification Lifecycle”Certifications are time-bound — they expire by design.
| Status | Meaning |
|---|---|
| 🟢 Certified | Meets all required policies |
| 🟡 Certified with Observations | Acceptable but with minor issues |
| 🟠 Probationary | Requires improvement soon |
| 🔴 Decertified | Fails required policies |
| ⚪ Expired | Certification window elapsed — needs recertification |
Default certification window: 90 days. Risk factors adjust automatically — high-churn code gets shorter windows, stable code gets longer ones.
Architect Review
Section titled “Architect Review”The certify architect command runs a 6-phase AI analysis of your entire codebase:
- Architecture narration — how packages compose, where data flows
- Code quality patterns — anti-patterns, complexity hotspots
- Test strategy — coverage gaps, missing test categories
- Security & operations — structural risk from AST analysis
- Recommendations — specific improvements with projected metric deltas
- Synthesis — executive summary, risk matrix, prioritized roadmap
All metrics are grounded in real AST-computed data — no fabricated values.
Workspace Mode
Section titled “Workspace Mode”For multi-repo projects using git submodules, --workspace mode treats your entire workspace as a single system:
certify certify --workspace # certify all submodulescertify architect --workspace # holistic system-level reviewcertify report --workspace # aggregated workspace reportCross-submodule dependencies, shared library quality, and system-level architecture are analyzed holistically.
Report Card
Section titled “Report Card”A complete per-unit certification of your entire codebase:
# 🟢 Certify — Report Card
## Dimension Averages| correctness | 95.0% | ██████████████████░░ || maintainability | 93.3% | ██████████████████░░ || readability | 92.4% | ██████████████████░░ || testability | 90.0% | █████████████████░░░ |
## All Units| Score | function | B+ | 87.4% | certified | 2026-06-07 || CertifyUnit | function | B | 85.6% | certified | 2026-06-07 |... every unit in your repoAlso available as a browsable static HTML site: certify report --site
Editor Integration
Section titled “Editor Integration”VSCode Extension
See certification grades inline on every function. Open an interactive dashboard. Configure AI providers visually. Browse models from any OpenAI-compatible endpoint.
code --install-extension iksnae.certify-vscodeSupports 11 provider presets — OpenRouter, Groq, Together, Fireworks, OpenAI, Google AI, Ollama, LM Studio, vLLM, and any custom endpoint.
Ready?
Section titled “Ready?”Install in 30 seconds
go install github.com/iksnae/code-certification/cmd/certify@latestcertify onboardCertify your first repo
Walk through bootstrapping, scanning, certifying, and reading your first report card. Use certify onboard to track your progress.