Skip to content

Lint the code your AI agent writes

Lint AI-generated code and install rules for AI coding agents.

bun add --dev --exact @gspothq/clibunx --no-install gspot init
Read the docs

Runs the tools your code needs

  • ESLint
  • Prettier
  • Ruff
  • Stylelint
  • ShellCheck
  • Knip
  • Gitleaks
  • OSV-Scanner
  • Trivy

Caught in gspot's own code

A coding agent added a package that brought in a known security hole. gspot stopped the push.

  1. The agent added

    docs/package.json
    "dependencies": {"mermaid-isomorphic": "3.1.0"}

    It pulls in katex 0.16.47.

  2. gspot stopped the push

    $ git push

    dependencies/osv-scanner

    bun.lock · GHSA-238p-pmpm-9mq7

    katex 0.16.47: Existing prototype pollution can bypass trust restrictions

  3. The fix

    package.json
    "overrides": {"katex": "0.18.11"}

    katex 0.18.2 and later fix the advisory.

What you get

  • Catch problems before every commit

    When you commit, gspot checks the files you changed, lists each problem, and stops the commit.

  • Set up all your linters with one command

    Run gspot init. It finds the languages and frameworks in your repository, turns on linters and formatters for them, and writes their settings.

  • Run the same tool versions everywhere

    gspot pins each tool that it installs, so your machine, your teammates' machines, and CI run the same versions.

  • Give your coding agent the rules

    gspot adds rules to AGENTS.md that tell coding agents how to write code that passes the checks.

  • Keep your formatting settings

    gspot init copies the indentation, line width, and quote style from your Prettier, EditorConfig, or Ruff settings. With none of those, it uses the defaults of each formatter.

  • Keep secrets and vulnerable packages out

    A commit that adds an API key or a token stops. A push that changes a lockfile stops when a locked package has a known security hole.

See all checks