Awesome MCP › Software Development

hyperb1iss/lucidity-mcp

⭐ 90 Python added to this list on 2025-05-06 repository created 2025-03-19

Lucidity MCP improves the quality of AI-generated code by giving the reviewing model structure rather than by running a linter. It reads the changes in a git working tree or diff and issues prompts that walk the model through defined analysis dimensions — unnecessary complexity, security problems, error handling, duplicated logic and the other categories where generated code characteristically goes wrong. The approach is prompt-based and language-agnostic: there is no per-language rule set, so the same server applies to any codebase, and the analysis is performed by the model under guidance rather than by static rules that must be written and maintained. The design point is that an assistant asked simply to review a diff tends to comment on surface style, while the same assistant given an explicit checklist finds substantive issues. Because it works from git changes it is naturally scoped to what was just written, rather than reporting on an entire repository. It is a Python project built on FastMCP, requires Python 3.13 or newer, is Apache-2.0 licensed, and uses ruff and mypy in its own development. The README describes it as in active development. The audience is developers using coding agents who want the agent's own output reviewed before it is committed.

https://github.com/hyperb1iss/lucidity-mcp

mcpcode-qualitycode-reviewpythonfastmcpgit-analysisstatic-analysisprompts