ai-code-quality 开源项目推荐（15 个）

sudokar /openspec-plus

OpenSpec Plus——通过更好的发现、需求、设计决策、执行规划与执行，增强 OpenSpec 规范驱动开发的 Agent 技能。可与 Claude Code、OpenCode、Github Copilot 及任何其他 AI 编码 Agent 配合使用

flamehaven01 /AI-SLOP-Detector

别再发AI垃圾了。检测空函数、伪造文档和AI生成代码中夸大的注释。准备生产。

bhvbhushan /vibecop

AI code quality toolkit — deterministic linter for the AI coding era. 22 detectors, GitHub Action PR gate, zero LLM required.

randomittin /heimdall

Verification gates for AI coding agents — nothing ships unproven. Falsifiable oracle gates, team wall, receipts. Works with Claude Code today; Cursor, Codex, Gemini CLI & every git repo next.

umitkacar /ai-coding-patterns

Production-ready AI/ML code patterns for Claude, GPT & Gemini - 590 Python snippets, 264 Mermaid diagrams, 99.3% quality with LLM-optimized context

NeelPrime /arbiterx

Your AI writes code. ArbiterX decides if it's good enough. Intelligent middleware for AI coding assistants — 97% token reduction, smart model routing, quality gate scoring. Works with Claude Code, Codex, Cursor, Copilot, Aider & more.

mohamedgame /stop-slop

🛠️ Remove AI patterns from writing with Stop Slop, a skill designed to guide Claude and other LLMs towards clearer, more authentic prose.

PraveenJayaprakash-JP /vibe-gate

Quality gate for AI-generated web apps. One command scans for security holes, broken flows, and auth gaps.

Alok-Fusion /ship-risk

AI-code readiness scanner — flags what a human reviewer would catch before AI-generated code ships (secrets, missing auth, weak validation, no tests) with explainable, weighted scoring.

mataluni-bravetto /biasguard-production-suite

🛡️ 17KB justice-driven AI bias detection suite born from fighting for Jahmere Webb's freedom. Real-time bias analysis for chat interfaces, development environments & codebase mapping. <10ms speed, 84.9% accuracy, zero dependencies. Mission: 15% revenue supports criminal justice reform.

Ashutosh-Panda2004 /DriftLens

An observability platform for AI-assisted development that measures ROI in dollars, predicts failures before they happen, and routes code to the best AI agent

balyakin /archlint

Enforce architectural boundaries in Python, TypeScript, and JavaScript codebases with deterministic YAML rules.

PeppaPigw /Slopsentinel

该项目在 GitHub 上没有填写描述

jwmurray /human-readable-code

A skill that keeps AI-generated code at human scale — readable, maintainable, and debuggable by a person. Applies at design, coding, and review.

vibesafeio /vibecost-action

AI writes code that works. Not code that's fast. VibeCost catches inefficient patterns in PRs.