Registry now open

Certified skills.
Behavioral contracts.
Zero trust by design.

The certification registry for AI agent skills. Every SKILL.md package gets a behavioral contract, an automated test suite, a CLR-ID certificate, and an x402 micropayment gate. Powered by Base. Expert review flags every mismatch.

Get certified -- from $29 See how it works
CLR-IDPer-package certificate
x402Native micropayment gate
3-layerFirewall isolation
On-chainEnterprise+ registry
How it works

From SKILL.md to certified package.

Every skill passes through a four-stage pipeline. The firewall layer enforces zero context bleed between Spec Team and Build Team LLMs.

01 — Ingest

SKILL.md Parse

The submitted SKILL.md is parsed into a structured behavioral contract — trigger conditions, capabilities, tool calls, and I/O schema. No implementation included.

Behavioral Contract
02 — Isolate

Firewall Layer

The Spec Team LLM context is hermetically sealed. The Build Team LLM receives only the behavioral contract — never the original SKILL.md source.

Firewall Layer
03 — Test

Behavioral Test Suite

An automated test suite runs the rebuilt skill against every trigger condition in the behavioral contract. Any mismatch raises an expert review flag immediately.

Expert Review Flag
04 — Certify

CLR-ID Issued

On full pass, a CLR-ID certificate is issued — a unique base-36 identifier binding the skill package to its behavioral contract, test results, and issuer signature.

CLR-ID
Certificate registry

Every skill, one certificate.

CLR-IDs are immutable records binding a skill package to its behavioral contract and test results. Enterprise+ tier mirrors every cert to Base (L2) for on-chain verification.

CLR-4X9KZ2M
CERTIFIED
skill_nameweb-search-v2
issued_at2026-03-28T00:00:00Z
behavioral_contractbc_sha256:a3f9d2...
test_suite_pass48/48 ✓
expert_reviewCLEARED
firewall_layerVERIFIED
on_chain_mirrorPowered by Base — 0x4f9d...
payment_gatex402 · $0.0012/call
⚑

Expert review flag on mismatch

Any behavioral deviation between the original contract and the rebuilt skill halts certification and routes to expert review. No cert is issued until cleared.

⛓

On-chain mirror for Enterprise+

PostgreSQL is the primary store for Dev and Team tiers. Enterprise+ certs are additionally mirrored to Base (L2) — tamper-proof, publicly verifiable, forever.

⚡

x402 micropayment gate

Every certified skill is wrapped with an x402 payment gate. Callers pay per invocation. Float is held between collection and inference provider settlement via OpenRouter.

Credits & Pricing

Credit-based. No subscriptions.

1 Skill Credit = 1 Standard certification. Enterprise certification uses 5 credits (includes on-chain mirror). Buy credits via x402 (Powered by Base, USDC). All tiers include behavioral contracts, automated test suites, and CLR-ID certificates.

🎁 First cert free — claim your trial credit at GET /api/trial-cert?email= · No credit card required.

Solo

1 Credit

$29
$29 per credit · pay-as-you-go
  • ✓1 Standard certification
  • ✓Behavioral contract generation
  • ✓Automated test suite (48 checks)
  • ✓CLR-ID certificate
  • ✓Expert review on mismatch
  • ✓Pay via x402 (USDC on Base)
Get certified
Team

5 Credits

$129
$25.80 per credit · 11% savings
  • ✓5 Standard certifications
  • ✓Everything in Solo
  • ✓Team credit code sharing
  • ✓Multi-agent firewall validation
  • ✓Priority expert review queue
  • ✓Volume savings over pay-as-you-go
Purchase credits
Studio

20 Credits

$449
$22.45 per credit · 23% savings
  • ✓20 Standard certifications
  • ✓Everything in Team
  • ✓Continuous behavioral regression
  • ✓OpenRouter model arbitrage
  • ✓Bulk certification pipeline
  • ✓Best value for teams at scale
Purchase credits
Enterprise

Enterprise Certification — $199

Uses 5 Skill Credits. Includes on-chain cert mirror. Powered by Base — tamper-proof, publicly verifiable.

Get certified
Platform features

Built for production agent systems.

Every design decision optimises for correctness, isolation, and verifiability at scale.

📋

Behavioral Contract JSON Schema

Trigger conditions, capabilities, tool calls, and I/O schemas captured as a formal JSON contract. The contract — never the source — is the unit of trust.

🚩

Automated mismatch detection

Test suite runs the rebuilt skill against every declared behavior. Divergence from the behavioral contract triggers an immediate expert review flag before any cert is issued.

🔒

Firewall layer enforcement

Spec Team and Build Team LLM contexts are hermetically isolated. Zero source code exposure during reconstruction — the same legal protection as semiconductor cleanrooms.

⛓

On-chain cert registry — Powered by Base

Enterprise+ certs are mirrored on-chain. Powered by Base — immutable, publicly verifiable, independent of any single infrastructure provider. Eliminates registry as a single point of failure.

⚡

x402 micropayment gate

HTTP x402 Payment Required as an API-native payment gate. Callers pay per invocation. Skill publishers receive float proceeds after OpenRouter inference settlement.

🛡

Sovereign Tier deployment

Air-gapped, self-hosted deployment of the full CleanSkills pipeline for regulated or classified environments. No cloud dependency. No external cert authority required.

🔍

Verification API — real-time trust checks

GET /api/verify/{clr_id} returns certified status, last tested timestamp, and an x402_trust signal. One call tells any agent whether a skill is safe to invoke.

📡

Webhook callbacks

Pass a webhook_url when submitting via POST /agent/certify. We fire certification.passed or certification.failed the moment review completes — no polling required.

📄

OpenAPI spec — auto-generate your client

Full OpenAPI 3.0 spec at /api/openapi.json. Every endpoint, every request and response schema, security schemes for x402 and credit codes. Drop it into any agent framework.

🎁

First cert free

Claim a free trial credit at GET /api/trial-cert?email=. One per email. Validate the full pipeline — behavioral contract, 48-point test suite, CLR-ID — before spending a dollar.

Glossary

Terms of the trade.

Precise definitions for every concept in the CleanSkills platform.

Clean Room
A documented legal methodology for producing an independent implementation with zero exposure to the original source. Validated by courts since Baker v. Selden (1879).
SKILL.md
The standard file format for agent skills in Claude Code, Codex CLI, and compatible agents. The unit of submission to the CleanSkills registry.
Behavioral Contract
A JSON schema capturing trigger conditions, capabilities, tool calls, and I/O schema — but no implementation. The firewall between Spec and Build teams.
Firewall Layer
The enforced separation between the Spec Team LLM context and the Build Team LLM context. Neither context can read from the other's window.
x402
HTTP x402 Payment Required used as an API-native payment gate. Per-call billing at sub-cent granularity, settled on-chain via Base (USDC).
CLR-ID
Certificate ID — a unique base-36 identifier for each certified skill package. Binds the skill to its behavioral contract, test results, and issuer signature.
Float
Working capital held between customer payment collection and inference provider payment. Managed via OpenRouter to minimise direct provider exposure.
OpenRouter
LLM inference aggregator used as the intermediary to minimise direct provider float and enable model arbitrage across the inference market.
Sovereign Tier
Air-gapped, self-hosted deployment of the full CleanSkills pipeline for regulated or classified environments. No external registry dependency.
Expert Review Flag
Triggered automatically when a rebuilt skill's behavior diverges from its behavioral contract during the test suite. Certification is halted until an expert reviewer clears the flag.
Start now

Certify your first skill.

Submit a SKILL.md package for certification. Solo tier starts at $29 per skill. Credits never expire.

Submit for Certification Buy Credit Bundles

Or subscribe for updates: