A GitHub Action that runs Claude Code on PRs and issues to answer questions, review changes, and implement code.
kunchenguid/no-mistakes
A local git proxy that runs AI-driven validation on your branch and opens a clean PR only after every check passes.
- Stars
- 9,159
- Forks
- 956
- Open issues
- 244
- Last commit
- 22 hours ago
What is it?
no-mistakes sits between your machine and your real git remote. Instead of pushing to origin, you push to a local no-mistakes remote. It creates a disposable worktree for the branch, runs a pipeline of review, test, docs, lint and push steps driven by a coding agent, and forwards the branch to the configured push target only after every local gate passes. It then opens a pull request automatically.
Each step either passes or stops with a finding. Safe mechanical fixes are applied automatically, while changes that affect intent are escalated for you to approve, fix or skip, through a TUI or a non-interactive interface. When CI fails, the pipeline can repair it and publish the fix through a guarded force-push, or send the repair back through review when it cannot confirm it builds on the reviewed head.
The project supports several agents, including claude, codex, opencode, copilot and cursor, with ordered fallbacks. It also installs a /no-mistakes skill so a coding agent can complete a task and run it through the gate.
Key features
- 01Runs review, test, docs, lint and push checks in an isolated worktree.
- 02Forwards the branch to the push target only after all checks pass.
- 03Opens a pull request automatically and watches CI.
- 04Auto-fixes safe issues and escalates the rest for human approval.
- 05Supports multiple coding agents with ordered fallbacks.
- 06Includes a /no-mistakes skill for coding agents.
Use cases
- Gating AI-generated changes before they reach a shared repository.
- Pushing a feature branch and getting a clean, reviewed PR without manual steps.
- Having a coding agent finish a task and pass it through automated checks.
- Contributing from a GitHub fork while keeping origin pointed at the parent repo.
Quick start
- 1
Run the official install script from the README
- 2
Initialize the gate in your repository
no-mistakes init - 3
Push your branch through the gate
git push no-mistakes - 4
Open the TUI to review findings for the active run
no-mistakes
Things to know
- Every gate requires a runnable configured AI agent, such as claude, codex, or opencode.
- Running the pipeline spends AI agent API quota.
In collections
Topics
FAQ
- What is no-mistakes?
- A local git proxy that runs AI-driven validation on your branch and opens a clean PR only after every check passes. no-mistakes sits between your machine and your real git remote. Instead of pushing to origin, you push to a local no-mistakes remote. It creates a disposable worktree for the branch, runs a pipeline of review, test, docs, lint and push steps driven by a coding agent, and forwards the branch to the configured push target only after every local gate passes. It then opens a pull request automatically.
- Is no-mistakes free to use? What license does it have?
- no-mistakes is open source under the MIT license. Check the license terms before using it in commercial products.
- Is no-mistakes still maintained?
- Actively maintained. The last commit to no-mistakes was on Nov 7, 2026.
- What can I use no-mistakes for?
- Gating AI-generated changes before they reach a shared repository. Pushing a feature branch and getting a clean, reviewed PR without manual steps. Having a coding agent finish a task and pass it through automated checks. Contributing from a GitHub fork while keeping origin pointed at the parent repo.
- How do I install no-mistakes?
- Run the official install script from the README Initialize the gate in your repository: no-mistakes init Push your branch through the gate: git push no-mistakes Open the TUI to review findings for the active run: no-mistakes
- Where can I get no-mistakes?
- The source code and installation instructions are on GitHub: https://github.com/kunchenguid/no-mistakes
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Automated dependency update tool that scans repos for outdated packages and opens pull requests to upgrade them.
Ruby tool that runs in CI to automate pull request review rules, such as enforcing changelogs and PR description conventions.
Danger JS runs after your CI to automate code review conventions, such as enforcing changelogs or PR description rules, in pull requests.
Draws each pull request as animated architecture and data-flow diagrams, posted as a comment inside the PR, for GitHub, GitLab and Bitbucket.
A CLI and GitHub Action that runs after AI code review to score a pull request's merge risk and list the files and endpoints it reaches.