Skip to content

Compliance checks for the code your AI writes

Lintsure works out which compliance standards apply to you, then checks your code against them. Every finding shows the file and line, the rules it breaks, and a fix.

npx lintsure scan
Add to your agent
Lintsureclaude code · ~/acme-api
rules v0.10 bytes uploadedexample run

Example runs. One: Lintsure scans a repo and finds a user email in the logs, a committed secret key and a loan score with no human review. Two: it proposes a one-line fix, you approve it, and the re-check marks it Met. Three: it explains that loan scoring is high-risk under the EU AI Act from 2 December 2027. Four: it lists what's left before a SOC 2 audit.

A readiness tool, not an audit or legal advice. It tells you where you stand and what to fix next.

Lintsure reads your code for the questions an auditor will ask

Hover to pause. Click a question to see the framework behind it.

From one prompt to proof, in three steps

You ask your coding agent. Lintsure finds where personal data goes, checks each rule the same way every time, and hands back fixes for you to approve.

It runs inside the tools you already use, and every finding maps to every framework

(Runs locally)

Lintsure never uploads your code

↗

Scan · ~/acme-api local

  • Code uploadedNo
  • Code indexedNo
  • Network callsNone
  • Reportlintsure-report.md

One finding, every framework at once

Pick a finding. Lintsure shows every rule it breaks across all four frameworks, so you fix it once and close several gaps together.

EU AI ActNot affected
GDPRArt. 32 Security of processing
SOC 2CC6.1 Logical access
ISO 27001A.8.11 Data masking

Learn what the rules ask of your code

Plain guides for software teams, and a page for each of the 47 checks with an example and a fix.

Shipped in the open

Every release is public, with the commits behind it. Here's what changed lately.

Full changelog →
  1. v0.3.0 · 9 Nov 2026

    OWASP Top 10:2025

    A fifth framework in Beta: all ten OWASP Top 10:2025 categories, described in our own words, with 31 new rules and eleven existing rules mapped to them.

  2. v0.2.0 · 8 Nov 2026

    Reports you can act on, rules you can measure

    Lintsure now remembers what your team already knows: a baseline shows what is new, fixed or unchanged, and accepted findings carry a reason and a name. Every rule is scored against a labelled test set.

  3. v0.1.3 · 8 Nov 2026

    Ask first, every time

    A hard rule for every agent that uses Lintsure: explain what it plans to do, point to the exact line, say why and what you gain, then wait for a clear yes.

If you run Lintsure in your company, send a review

Good or bad, we read every one. We only publish it if you say we can.

Send a review

Common questions

Does Lintsure upload my code?

No. It runs on your machine or in your own CI and never sends your code to us. Your coding agent, if you use one, follows its own provider's rules.

Will this make me compliant?

No tool can do that alone. Lintsure is a readiness tool, not an audit or legal advice. It shows what your code proves, what it doesn't, and what lives outside code, like policies and training.

Why not just ask Claude Code?

A general agent will give a different answer each time, and it may say a control is met when it isn't. Lintsure's Checker is a fixed rule engine: same input, same result. Only it can mark a control as Met.

What does it cost?

The checker is free and open source. Lintsure Monitor, a paid service with checks on every pull request and review by a real auditor, is coming later. Request a demo to see it.

Read all questions →

Set up in a minute.
Check in one prompt.

Open source · Runs on your machine · Free

claude mcp add lintsure -- npx -y @lintsure/mcp
npx lintsure scan
uses: lintsure/action@v1