Is Your Repo
AI Agent-Ready?

Benchmark how well autonomous coding agents (Claude Code, Cursor, Codex, Devin) navigate and execute in your repository. Score from 0 to 100, turn missing context into ready-to-merge diffs, and show off a live README badge.

Run instantly from your terminal:

$ npx agentbadge scan . --badge --fix

The 5 Pillars of Agent Readiness

Repositories are no longer just for human eyes — they must expose machine-readable context.

1. Context & Boundaries (25 Pts)

AGENTS.md, CLAUDE.md, and documented architecture. Does the agent know its guardrails within 60 seconds?

2. Setup & Environment (20 Pts)

Deterministic lockfiles, .env.example templates, and containerized environments for one-command boots.

3. Verification & Feedback (20 Pts)

Can the agent test its own code in 30 seconds? Automated test suites, strict type-checking, and formatters.

4. Guardrails & Safety (20 Pts)

CI/CD automation, pre-commit hooks, and secret leak prevention to prevent broken code from shipping.

5. Tooling & Machine Schemas (15 Pts)

Model Context Protocol (MCP) tool manifests, OpenAPI/Prisma models, and machine-readable logs.

Is Your Repo AI Agent-Ready?