Kira
Automatically provides AI agents with proven instructions and past failure warnings for common tasks like deployment, auth, and payments, enabling flawless execution without manual configuration.
README
Kira
One MCP. Your agent becomes a genius.
Stop managing CLAUDE.md files, .cursorrules, and skill folders across projects. Install Kira once — your AI agent automatically finds the right instructions, avoids known mistakes, and executes flawlessly.
Privacy by design. Kira learns from agent outcomes via opt-in telemetry that redacts secrets, paths, and identifiers locally before write AND server-side before storage. Run
npm run demo:privacyto see exactly what leaves your machine. Full wire format and opt-out in PRIVACY.md.
Install (10 seconds)
Add this snippet to your MCP host config:
{
"mcpServers": {
"kira": {
"command": "npx",
"args": ["kira-mcp"]
}
}
}
That's it. Your agent now has Kira.
<details> <summary><b>Per-client paths (click)</b></summary>
| Client | Config file |
|---|---|
| Claude Code | ~/.claude/settings.json (global) or .claude/settings.json (per-project) |
| Claude Desktop | macOS: ~/Library/Application Support/Claude/claude_desktop_config.json · Windows: %APPDATA%\Claude\claude_desktop_config.json |
| Cursor | ~/.cursor/mcp.json (global) or .cursor/mcp.json (per-project) |
| Cline / Continue | extension settings → MCP servers |
| Windsurf | ~/.codeium/windsurf/mcp_config.json |
| VS Code (MCP preview) | .vscode/mcp.json |
| Goose | ~/.config/goose/profiles.yaml (under extensions:) |
| Zed | ~/.config/zed/settings.json (context_servers) |
The snippet above works as-is in every one of them — just paste it under mcpServers (or the equivalent key for your client).
</details>
Demo

What happens
Before Kira: Agent deploys to Vercel, forgets env vars, app crashes. Retries 3 times. Burns tokens.
After Kira: Agent automatically calls kira_lookup("deploy vercel") before acting. Gets step-by-step instructions + a Scar warning: "847 agents forgot env vars — run vercel env ls first." Deploys correctly on the first try.
Three tools, zero config
| Tool | What it does |
|---|---|
kira_lookup |
Give it a keyword ("stripe", "deploy", "auth") → get proven instructions + past failure warnings |
kira_route |
Give it a goal ("build a web app") → get an ordered plan with skills for each step |
kira_report |
Agent reports success/retry after each task → feeds the quality system |
Auto-firing
You don't call Kira. Kira tells your agent to call it. Via MCP instructions, the agent automatically looks up skills before starting any task. You literally do nothing.
What's inside
31 Skills (and growing daily)
| Category | Skills |
|---|---|
| Deploy | Vercel, Cloudflare Pages |
| Database | Prisma, Drizzle, Supabase |
| Auth | Clerk, Auth.js v5 |
| Payments | Stripe Checkout |
| UI | Tailwind CSS v4, shadcn/ui |
| Testing | Vitest, Playwright E2E |
| CI/CD | GitHub Actions |
| Infra | Docker, ESLint flat config |
| Services | Resend email, React Email, Sentry, tRPC, S3/R2 upload, Upstash Redis |
| Background | Inngest |
| State | Zustand, Zod validation |
| Upload | UploadThing, S3/R2 |
| Observability | PostHog analytics, Sentry |
| Mobile | Expo / React Native |
| i18n | next-intl |
| CMS | Payload CMS |
| Monorepo | Turborepo |
12 Scars (past failure patterns)
Scars warn your agent about mistakes other agents already made:
- Next.js "use client" directive missing — client hooks in server components
- Vercel deploy succeeds but app crashes — missing env vars
- Stripe webhook signature fails — body already parsed
- Auth.js signIn/signOut wrong import — server/client mixup
- Prisma types are stale — forgot
generate - Clerk middleware in wrong directory — auth silently broken
- Supabase RLS not enabled — data publicly exposed
- Tailwind v4 PostCSS config wrong — v3 plugin breaks v4
- Vitest path alias mismatch — tsconfig vs vitest.config desync
- And more — hit counts updated from real agent data
7 Routes (goal-to-plan)
Ask "build a web app" → Kira returns 8 ordered steps, each with its skill and scars:
1. Tailwind CSS v4
2. shadcn/ui
3. ESLint
4. Prisma + Scar: "don't forget prisma generate"
5. Clerk Auth + Scar: "middleware goes in root, not app/"
6. Vitest
7. GitHub Actions CI
8. Vercel Deploy + Scar: "check env vars before --prod"
How it works
Your agent gets a task
↓
Kira auto-fires (MCP instructions)
↓
kira_lookup("deploy vercel", context: ["nextjs"])
↓
Returns: Skill (step-by-step) + Scars (what to avoid)
↓
Agent announces choice → follows instructions → reports result
↓
kira_report("community.deploy-vercel-nextjs.v1", "success")
Skills are natural language Markdown — no executable code, no injection risk.
Why not just use CLAUDE.md?
| CLAUDE.md / .cursorrules | Kira | |
|---|---|---|
| Setup | Copy per project | Install once |
| Updates | Manual | Automatic |
| Selection | You choose | Agent chooses |
| Failure avoidance | None | Scars (past failures) |
| Multi-step planning | None | Routes |
| Quality tracking | None | success/retry scoring |
| Works across AI tools | Tool-specific | Any MCP client |
Telemetry
Kira sends anonymous outcome data to a central Worker so the community can improve Skills and surface new Scars.
Mode (KIRA_TELEMETRY env, or kira_consent MCP tool) |
What leaves your machine |
|---|---|
off |
Nothing. Local log only. |
basic (default) |
Anonymous core: skill ID, status, anonymous UUID, kira version, OS family, Node major version, free/pro tier. No free text. |
full |
Same as basic plus sanitized note / context (secrets, paths, identifiers redacted). |
Full schema, redaction rules, retention, and opt-out instructions: PRIVACY.md.
| Env var | Default | Purpose |
|---|---|---|
KIRA_TELEMETRY |
(unset → basic) |
Override consent level for this process: off, basic, full. |
KIRA_TELEMETRY_URL |
https://kira-telemetry.workers.dev/v1/reports |
Endpoint for batch upload. |
KIRA_HOME |
~/.kira |
Where consent state and the local log live. |
Contributing
The first 1,000 contributors get permanent free access to all Kira features (including future Pro tier).
See CONTRIBUTING.md for how to add Skills and Scars.
Links
Where agents shine.
A B Button Corporation project.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.