REPO REVIEWS

codebase-memory-mcp builds a persistent knowledge graph of your codebase and exposes it via MCP — indexing the Linux kernel in 3 minutes, answering structural queries in under 1ms. Here's an honest look at whether it lives up to the claims.

codebase-memory-mcp is a single static binary that indexes your entire codebase into a persistent knowledge graph, giving AI coding agents like Claude Code and Cursor structural query access in under 1ms. It's genuinely fast, surprisingly well-engineered, and worth a serious look if you're tired of your AI assistant reading files one by one.

Tavily's MCP server gives AI agents like Claude real-time web search, extraction, crawling, and site mapping through a single integration. It's one of the more polished MCP servers out there, but it comes with the usual tradeoffs of depending on a paid third-party API.

A massive open-source AI curriculum spanning 20 phases, 260+ lessons, and ~290 hours of content — from linear algebra to autonomous agent swarms. I dug into what's actually there, what's still vaporware, and whether it's worth bookmarking.

fireworks-tech-graph is a Claude Code skill that generates SVG and PNG technical diagrams from natural language descriptions, with 7 visual styles and deep AI/Agent domain knowledge baked in. It's four days old, already has 2800+ stars, and is worth a serious look — with some caveats.

GSD 2 is a standalone TypeScript CLI that wraps LLM coding agents with real session management, spec-driven execution, and autonomous milestone progression. It's genuinely ambitious, actively developed, and rough enough around the edges that you should know what you're signing up for.

executeautomation/mcp-playwright gives Claude, Copilot, and Cursor direct control over a real browser via the Model Context Protocol. With 5,400+ stars and active weekly commits, it's become the de facto standard for AI-driven browser automation — but there are some rough edges worth knowing about before you adopt it.

mcp-playwright connects Playwright to Claude, Cursor, and other MCP-compatible AI tools, letting LLMs actually drive a browser. It's one of the more practical MCP servers out there, but there are real tradeoffs worth understanding before you wire it into your workflow.

Humanizer is a Claude Code skill with nearly 14k stars that strips AI-writing patterns from text using a curated list of 29 specific anti-patterns. It's not magic, but it's more systematic than anything else I've seen in this space.

Vibe Kanban hit 25k stars by solving a real problem: managing multiple AI coding agents across a project without losing your mind. The catch? The company behind it, Bloop, is shutting down — and that changes everything about whether you should adopt it.