Nikita-Filonov/ai-review ? reverse-engineered prompt

Build me a Python command line tool that can automatically review code changes in pull requests and merge requests using an LLM.

I want it to work with GitHub, GitLab, Bitbucket Cloud, Bitbucket Server, Azure DevOps, and Gitea, and be able to post inline comments, a summary review, and replies inside existing discussion threads. It should support a few review modes, like full review, inline only, context aware review across multiple files, and reply mode for follow up comments.

Make it configurable from YAML, JSON, or environment variables, so I can point it at different AI providers like OpenAI, Claude, Gemini, Ollama, Bedrock, OpenRouter, or Azure OpenAI. It should run locally or in CI, read the repo it is reviewing, and use the provider and VCS API tokens from config. If useful, look up current docs online for any API details.

Please include a Docker setup and a simple way to run it from the command line like run, run-inline, and run-summary, so it feels easy to drop into a pipeline.