Where agentlint fits — and where it doesn't
agentlint measures one thing nothing else measures: agent-readiness — how well an AI coding agent can operate in your repo. Here's an honest map of the neighbours.
Codecov
Free for OSS; ~$10/user/mo ProTest-coverage tracking: CLI uploads coverage from CI, dashboard shows trends, PR comments, badges.
Use it when you need to track how much of your code your tests execute. Coverage is its whole job and it does it well.
Same product shape — CLI upload, dashboard, badge, PR comment — but the metric is agent-readiness, not coverage. The two are complementary, not substitutes.
SonarQube / CodeClimate / Codacy
~$10–17/user/mo hostedStatic analysis for bugs, vulnerabilities, and code smells inside your source code.
Use it when you need deep per-line static analysis for humans reviewing code quality and security.
agentlint doesn't read your source for bugs. It audits the repo around the code — docs, scripts, conventions, guardrails — the things an AI agent needs to work effectively in your repo.
Manual AGENTS.md checklists
FreeHand-maintained checklists and awesome-lists describing what a good agent-ready repo looks like.
Use it when you want to learn the reasoning behind each practice — the lists are great study material.
A checklist can't score 200 repos, fail a CI run, or notice the regression three months later. agentlint turns the checklist into an enforced, versioned, 0–100 metric with fix prompts attached.
Doing nothing
Free (visibly)Letting each agent rediscover your repo from scratch, every session.
Use it when AI agents never touch your codebase.
Every confused agent session costs tokens and review time. A 30-second scan plus an afternoon of prompt-driven fixes is usually the cheapest productivity win available.
The CLI is free. The dashboard is $5.
No per-seat math, no tiers to study. Scan locally forever, or pay the price of a coffee for trends, badges, PR comments, and policy gates across your whole org.