About

For the rule surface your agents follow.

Writing rules for an agent is the easy part. Knowing whether the agent reads them — and whether they contradict each other — is the hard part. Every other dev tool reviews the code your team writes. agentlint reviews the rules your agents follow.

What we build for

Indie developers and small teams who maintain a real rule surface — CLAUDE.md, AGENTS.md, .cursor/rules, SKILL.md, hooks, settings — and want to make their agent setup more reliable without writing their own linter.

We don't build for hobbyists with a one-line CLAUDE.md (no rule surface to audit) or for enterprises looking for a SOC2'd vendor-managed agent platform (different sales motion, different product). The middle is where the work is.

What's on the roadmap

What we promise to never do

Get in touch

Bug reports, false positives, severity miscalibration, harness coverage requests, anything else — team@agentlint.net. We read every email.