For vibe coders & devs shipping with AI

Your AI agent keeps breaking your app? Its instructions are the problem.

In 60 seconds we generate a rules kit tuned to your stack: AGENTS.md, CLAUDE.md, Cursor rules (.mdc), Copilot, Windsurf plus safety rails that stop the agent from deleting files or pushing to main.

  • ✅ Exact test, lint & build commands, so the agent stops guessing
  • ✅ "Always / Ask first / Never" rules
  • ✅ The mistakes agents repeat most in your stack
  • ✅ One canonical file instead of 5 drifting copies

Free preview, no signup. Full kit $7 one-time. · AI tools directory · Free AGENTS.md templates · Agent ignoring your rules? · Português

FAQ
What files do I get?

AGENTS.md (canonical), CLAUDE.md (@import), .cursor/rules/*.mdc, .github/copilot-instructions.md, .windsurfrules, plus Claude Code permission settings. Free preview shows ~40% of AGENTS.md.

Does this work with Cursor, Claude Code and Copilot?

Yes. Those tools (plus Codex, OpenCode, Windsurf/Devin Desktop) read AGENTS.md or the matching rules file. We generate the formats you pick.

Is .cursorrules still supported?

.cursorrules is legacy. We generate .cursor/rules/*.mdc with globs and keep shared rules in AGENTS.md so your context window stays small.

What if the kit doesn't fit?

Reply to the receipt within 7 days for a refund. One-time $7 — no subscription.

📬 Not ready? Get the free starter AGENTS.md template

Plus one short email a week on AI agents & vibe coding. Unsubscribe anytime.

1 of 4

Which agents do you use?

Pick all that apply.

CursorClaude CodeGitHub CopilotDevin Desktop/WindsurfCodexGemini CLIGrokLovable/Bolt

2 of 4

What's your stack?

Next.jsReact + ViteNode/ExpressNestJSTypeScriptTailwindSupabasePrismaPythonFastAPIDjangoRailsLaravelFlutterReact Native/ExpoGoRustUnity/C#GodotFirebasePostgresDocker
npmpnpmyarnbunpip/uvcargoother

3 of 4

About the project

SaaS / web appLanding / siteAPI / backendMobile appBot / automation / AI agentGame
Vibe coder (new to code)IntermediateSenior

4 of 4

How should the agent work?

Small changes, one step at a timeExplain before editingNever delete files without askingNever push to mainDon't add libraries without askingStrict typingMatch existing style
Always run tests before finishingWrite tests for new codeNo tests yet

Building your kit…

  • Reading your stack
  • Inferring build/test commands
  • Writing Always / Ask / Never rules
  • Mapping common agent mistakes
  • Generating per-tool versions
  • Scoring quality

Free preview

Your kit for is ready

Quality score
AGENTS.md
🔒 Rest unlocked in the full kit






$19 $7
One-time payment · card / Apple Pay / Google Pay · instant download




Guarantee: if the kit doesn't fit your project, reply to the receipt within 7 days for a refund.