mattzcarey/shippie
Your CI pipeline now has opinions about your code
Shippie drops an LLM reviewer into your CI pipeline to catch secrets, bugs, and slow code before they reach production.
Not currently ranked — collecting fresh signals.
star history
What it does
Shippie is an LLM-powered code review agent that plugs into CI/CD pipelines to flag exposed secrets, inefficient code, and unhandled edge cases before a merge. It also runs locally as a CLI tool to review staged files. The project is built in TypeScript and aims to act like a human reviewer armed with a small, optimised toolset.
The interesting bit
Where it diverges from a static linter is its role as a Model Context Protocol client, allowing it to pull in external context from tools like browser automation, infrastructure deployments, or observability dashboards while reviewing a diff. That turns code review from pattern matching into something more like investigation.
Key highlights
- CI/CD integration plus local CLI reviews of staged files
- Flags secrets, performance issues, and potential bugs via LLM analysis
- MCP client support for external tool integration (browser use, infra, monitoring)
- Configurable rules files and pluggable AI providers
- TypeScript codebase running on Bun
Verdict
Worth a look if you are already experimenting with AI-assisted review and want an extensible, pipeline-native agent. If you treat code review as a strictly deterministic process, an LLM-driven tool is unlikely to fit.
Frequently asked
- What is mattzcarey/shippie?
- Shippie drops an LLM reviewer into your CI pipeline to catch secrets, bugs, and slow code before they reach production.
- Is shippie open source?
- Yes — mattzcarey/shippie is open source, released under the MIT license.
- What language is shippie written in?
- mattzcarey/shippie is primarily written in TypeScript.
- How popular is shippie?
- mattzcarey/shippie has 2.5k stars on GitHub.
- Where can I find shippie?
- mattzcarey/shippie is on GitHub at https://github.com/mattzcarey/shippie.