Build production-ready software with AI

Enterprise Skills reviews what your AI writes. It checks the architecture, audits security, closes test gaps, and records proof that the release passed.

Start buildingSee how it works

Free plan available · No credit card required · 98 engineering skills in the catalog

No engineering background? The guided path sets everything up for you →

Works with the tools you already use

ClaudeGPT-5 / CodexGeminiCursorWindsurfVS CodeJetBrainsSee all supported tools

The difference is what happens after the AI writes code

Generation was never the hard part. Everything between generated code and a production deploy is, and that layer is the product.

Without Enterprise Skills

  1. Idea
  2. Prompt
  3. AI writes the code
  4. Manual review, if there's time
  5. Hope it works
  6. Production bugs find your customers

With Enterprise Skills

  1. Idea
  2. Prompt
  3. AI writes the code
  4. Architecture & code review
  5. Security audit
  6. Test & coverage gaps closed
  7. Documentation & changelog
  8. Proof recorded for every check
  9. Release Governor decides, and can say no
  10. Production deploy, verified

This page practices what it sells: 886 release decisions across 9 repositories, 416 verifiable ledger entries, counted from the ledger when this page was last rendered. See a real decision below →

886
release decisions
54
FAIL verdicts from the gate
9
repositories with recorded decisions
2,927
skill runs

Counted from our own release records when this page was last rendered. See all the numbers and their sources

See it work

One feature, two verdicts, and the evidence in between

A login feature went through the real product. The gate refused it, nine findings were fixed and two accepted on the record, and then it passed. Both decisions replay from a clone.

The repository is ours and we built it for this: an AI assistant wrote an unfinished login feature, then the same assistant reviewed it under the catalog skills. It is not a customer's code, and the gaps are the point.

$ enterprise-skills v4.22.1

$ enterprise-skills preflight

"Review the login feature before it ships"
Preflight resolved:
- Task type: implementation
- Model: sonnet (implementation)
- Skills available: (category match (quality))
    code-review [active] category:quality
    full-stack-audit [active] category:quality
- New skill needed: no
- Authority pack: loaded

Given the task in plain English, the governance layer resolves what kind of change this is and which model should do it. Same repository, same answer, every time. Output shown unedited.

files in this change

app/login/page.tsxlib/auth.tslib/session.tsapp/api/login/route.ts · added by the fixlib/audit-log.ts · added by the fix

The files this change touches, on a branch you can check out. The dashed files arrived with the fix, one of them because the finding was that it did not exist.

skills that recorded evidence

api-contract-checkcode-reviewdependency-auditload-and-resilienceobservability-qapre-deploy-checksecurity-audittest-gap-analyzer

Preflight names the skills matching this task's category; the governor's required evidence areas decide which ones must actually record evidence, which is why more appear here. A required area with no evidence blocks the merge.

11 findings · FAIL at 60561e737258

  • CRITICALapp/login/page.tsx:22api-contract-check

    The form posts to /api/login, which does not exist in this repository

  • CRITICALlib/session.ts:30security-audit

    Session cookie is issued without Secure, HttpOnly, or SameSite

  • HIGHlib/auth.ts:74observability-qa

    Sign-in outcomes were not recorded anywhere

  • HIGHlib/auth.ts:33test-gap-analyzer

    No test covers the login flow

Show the other 7 findings
  • HIGHlib/auth.ts:33security-audit

    No limit on failed credential attempts

  • HIGHlib/auth.ts:26security-audit

    Passwords hashed with unsalted SHA-256

  • HIGHpackage.json:8test-gap-analyzer

    The test script exits 0 without running a test

  • MEDIUMlib/auth.ts:30code-review

    Email is lowercased at lookup but never at storage

  • MEDIUMlib/auth.ts:36security-audit

    Hash comparison is not constant-time

  • MEDIUMlib/auth.ts:60load-and-resilience

    The attempt ledger is per-process, so the lockout scales with instances

  • MEDIUMlib/session.ts:16code-review

    Sessions are held in a process-local Map

=== Release Governor — FAIL ===
  commit  60561e737258
  score   0/100
  ! api evidence reports FAILED (api-contract-check)
  ! P0 in security (security-audit): session-cookie-flags
  ! P1 in security (security-audit): password-hash-unsalted
  ! P1 in security (security-audit): no-attempt-limiting
  ! security evidence reports FAILED (security-audit)
exit 1

Every finding is pinned to a file and a line. The governor returned FAIL and exited 1, the exit that stops a merge in CI, and named its reasons.

$ enterprise-skills govern

=== Release Governor — PASS ===
  commit  57399589ddda
  code              ✓
  service           ✓
  api               ✓
  security          ✓
  release-readiness ✓
  score   100/100 · PASS
exit 0

12 tests · 12 passed · 0 failed (node:test) · 9 fixed · 2 accepted

An accepted finding carries no score penalty once it names what tracks it. That is why this reaches 100 with two gaps still open by design.

The decision after the fixes: every required area answered, and a score anyone can recompute from the same evidence.

Captured 2026-08-09 from real runs. CLI v4.22.1 · FAIL at 60561e737258 · PASS at 57399589ddda. Clone it and replay the gate over the recorded evidence →

What reproduces exactly, and what varies

Replaying the governor over the committed evidence reproduces these verdicts exactly. Re-running the skills produces fresh findings whose wording, and sometimes count, will differ. The findings come from an AI agent following the catalog skills, not from a fixed analyzer.

These two decisions were run locally, so they are reproducible rather than countersigned. Decisions posted from CI are signed and countersigned. see a real, signed decision below

Built for wherever you're starting

Pick the description that sounds like you. Same product, tuned to your journey.

Build production-ready software without years of engineering experience.

  • A guided path that sets everything up for you
  • Ready-made engineering skills, no need to know what to ask for
  • Every change reviewed, with the results explained in plain language
  • A help center and glossary for every unfamiliar term
Start with the guided pathNew to the terminology? Learn the words first →Coming from Lovable, Bolt, or another app builder? →

What you get

98 engineering skills working in the background, so every change meets a professional bar.

Ship with confidence

Every change is reviewed against professional engineering standards before it lands, not after something breaks.

How it works, in depth
For developers
Catalog skills run architecture and code review on every change, in your IDE or CI; findings arrive as fixes to apply, not lint noise.
At enterprise scale
The Release Governor turns the same reviews into a merge gate that fails closed on missing or stale evidence.

Catch problems early

Security gaps, missing tests, and reliability risks surface while they're still cheap to fix.

How it works, in depth
For developers
security-audit, test-gap-analyzer, and dependency checks run against the actual diff, before human review.
At enterprise scale
Findings become recorded evidence under a severity policy: a critical can never be silently accepted.

Code you can grow with

Guidance that keeps your codebase maintainable as your product and your team get bigger.

How it works, in depth
For developers
Refactoring, documentation, and changelog skills keep structure and docs moving with the code.
At enterprise scale
One standard across every repo and every AI surface your team uses: nine development surfaces, same rules.

Proof for whoever asks

Every release decision is recorded and independently verifiable, for teammates, customers, or auditors.

How it works, in depth
For developers
Every release decision binds to the exact reviewed content: a digest, not a promise.
At enterprise scale
Your auditors can check the record themselves, without asking you for it. Each decision is signed and appended to a log that cannot be edited after the fact.

Works with your stack

One canonical library, rendered natively into your editors, your coding agents, and whatever model you're driving. No extension conflicts, no lock-in.

New here? Just run enterprise-skills init : it auto-detects your editor(s) and configures them. The commands below are only for forcing a specific one.

Your AI assistant

Generates software

  • Writes code from your prompts
  • Autocompletes, explains, and edits
  • Moves as fast as you can describe

Enterprise Skills

Makes it production-ready

  • Reviews architecture, security, and tests on every change
  • Keeps documentation and changelogs honest
  • Gates releases on verifiable evidence and records the proof

IDEs & editors

Agent apps & CLIs

Cursor

Free+

Native skills + rules via .cursor/skills/ and .cursor/rules/

enterprise-skills init --ide cursor

Your model

Auto-matched

The same standards, conditioned per model from each vendor's own guidance.

Anthropic
Claude Opus 4.8 · Claude Opus 4.7 · Claude Opus 4.5 · Claude Sonnet 4.6 · Claude Sonnet 4.5 · Claude Haiku 4.5
OpenAI
GPT-5 · GPT-5 Codex · Codex 5.6 Sol · o-series (reasoning)
Google
Gemini

Nothing to configure. A constant operating posture stays fixed across every model; per-model conditioning swaps automatically with the model you're on.

Under the hood

For engineers who want to see the machinery

The same evidence always produces the same decision, so a release is never a judgement call. Every decision is signed, and anyone can verify one, including you.

hub PR #85: mergedPASS
7130db519 filesdependency
  • codecode-review · dependency-audit · test-gap-analyzer
  • securitysecurity-audit

A dependency-labeled change with code + security evidence on disk: PASS, merge unblocked.

hub PR #84: blocked until evidence landedFAIL
1ac88c6a4 filesauth
  • codecode-review · dependency-audit · test-gap-analyzer
  • apimissing
  • securitysecurity-audit

required domain 'api' has no evidence (policy: on_missing=fail)

A change classified as needing api evidence that had none: FAIL, closed by policy, until the evidence landed.

The record behind that PASS

Countersigned
Repository
mawebb001/cursor-enterprise-skills
Commit
7130db516034
Decision
PASS · 100/100
Verify offline
Verify a record yourself

Unedited lines from the release-governor check on our own pull requests #85 and #84 (a few metadata lines omitted for width).

How release gates workVerification methodology

Simple, transparent pricing

Start free with eight core skills. Upgrade when you need the full catalog, release gates, more IDEs, or team governance.

Community

Eight core skills to start building well: free forever, runs in Cursor. Release gates and the CLI arrive with Pro.

Learn the workflow

$0/month
Start free
  • 8 included skills
  • 1 IDE integration
  • Release gates arrive with Pro
  • Community support: docs, glossary, and the guided path
Most Popular

Pro

Every standard skill, CLI tool, 4 adapters, standard governance, and skill compiler. For professional developers and freelancers.

Ship real products

$29/month/seat
Start Pro
  • 94 included skills
  • 4 IDE integrations
  • Release gates: included
  • Support: 48h response

Team

Everything in Pro plus all 8 IDE adapters, team governance dashboard, cross-project intelligence, execution logging, feedback capture, and the customer Evidence Center with 90-day retention. For engineering teams of 5-50.

One standard for the team

$49/month/seat
Start Team
  • 94 included skills
  • 8 IDE integrations
  • Release gates: with policy controls
  • Support: 24h response
  • Minimum 5 seats

Enterprise

All skills, all governance features, all adapters, plus compliance mapping, incident response, IP scanning, SAML SSO and SCIM provisioning, org roles and audit, verified Evidence Center exports, one-year evidence retention, and custom skill development.

Governance at scale

Custom

Org contracts from $20k per year

Contact Sales
  • 98 included skills
  • 9 IDE integrations
  • Release gates: with policy controls and audit
  • Support: 4h response
  • Minimum 10 seats

Volume discounts available: 15% off for 6-20 seats, 25% off for 21-50 seats, 35% off for 51-100 seats.

Pro and Team include every skill except four enterprise-reserved ones. See the full plan comparison →

Common questions

Does this replace Cursor, Claude Code, or Copilot?

No, it completes them. Your assistant keeps generating the code; Enterprise Skills reviews, tests, documents, and gates what it generates. One command installs the skills into the tools you already use.

Does my code leave my machine?

The skills run in your IDE and your CI, against your repository, with your credentials. Sharing evidence with the hosted Evidence Center is opt-in, allowlisted, and redacted before upload, and you can preview exactly what would be sent.

How is this different from a linter or a CI pipeline?

Linters check syntax; CI runs whatever you wired. Enterprise Skills reviews the change itself: architecture, security, test gaps, documentation. It then binds the results to the exact content reviewed, so a release decision is evidence, not a green dot.

Can I verify the evidence myself?

Yes. Every release can produce an independently verifiable record showing what was reviewed and what passed, checkable without trusting this website. The verifier page walks you through it.

Is the free plan actually usable?

Yes. Real skills, the guided path, and the same review engine. Free, no credit card.

I've never shipped software before. Is this for me?

It was built so you don't have to know what to ask for. Start with the guided path: it sets everything up, reviews every change, and explains results in plain language. The learn section teaches the twenty words you actually need.

Start building better software today

Free to start: one command to set up, working in the tools you already use.

Start freeFollow the guided path

And after you ship: LaunchOps helps you launch and grow; see how