Pular para o conteúdo
Kitbay
Servidor MCP

open-code-review

Roda localmenteBUSL-1.1Atualizado em 16 de abr. de 202640Estrelas no GitHubPode alterar dados

Open Code Review scans code produced by AI assistants and is offered as a CLI and as a local MCP server started with npx.

Resumo

Open Code Review scans code produced by AI assistants and is offered as a CLI and as a local MCP server started with npx. It looks for imports of packages that do not exist, outdated API calls, contradictions across files, needless abstraction and insecure patterns, and grades the result against a threshold so it can fail a CI pipeline. Deeper scan levels add embedding-based analysis and an LLM pass, using a local Ollama model or any OpenAI-compatible endpoint. Through MCP the agent gets three tools: run a scan, apply AI-generated fixes, and explain a finding. The code is published under the Business Source License 1.1.

O que faz

  • Scan code for imports of packages that do not exist
  • Flag outdated APIs, needless abstraction and insecure patterns
  • Score code against a quality threshold for CI
  • Apply AI-generated fixes to findings
  • Explain an individual finding