Best Oss AI Skills & MCP Servers

Browser-based IDE + AI agent orchestration CLI. Run Claude Code in parallel across git worktrees, auto-approve safe tools with the Security Bouncer, and control long-running AI work from any device at app.mstro.app.

Query WHO Global Health Observatory data — 3,059 indicators across 194 member states with country, region, year, and sex filters via MCP. STDIO or Streamable HTTP.

MCP server that gives coding agents typed 0xArchive market data tools across Hyperliquid (perps, HIP-3, HIP-4, spot) and Lighter.xyz

Cross-agent permission policy specification for AI coding agents

Sync MCP servers, Skills, Agents & Commands across Claude Code, Cursor, OpenCode, Codex. One config, all tools.

Reduce Claude Code token usage by 70-90% with cross-session memory and four-layer optimization: semantic codebase indexing, PostToolUse tool compression, turn summarization, and session consolidation. MCP server powered by Ollama — zero API calls, runs en

Package Intelligence MCP server for AI agents. Stops hallucinated/malicious package installs across 19 ecosystems (npm, PyPI, Cargo, Go, Maven, NuGet, RubyGems, Composer, Pub, Hex, Swift, CocoaPods, CPAN, Hackage, CRAN, Conda, Homebrew, JSR, Julia). 22 to

Unified CLI to manage MCP servers across Claude Code, OpenCode, and more

MCP server for InsumerAPI — condition-based access infrastructure. Send a wallet and conditions, get a signed boolean across 37 chains. 27 tools: attestation, wallet trust profiles (49 checks across 27 chains incl. Solana, XRPL, Bitcoin, Tron, Stellar, Su

Model Context Protocol (MCP) server for AI agents to manage ad campaigns across Google, Meta, LinkedIn, Microsoft, Reddit, TikTok, and more. Create campaigns, adjust budgets, generate creatives, and pull performance data through natural conversation.

Resolve DOIs, search ~155M scholarly works, and fetch references via the Crossref REST API. STDIO or Streamable HTTP.

Give Claude Code superpowers — parallel multi-model queries, smart routing, and image/video generation across OpenAI (GPT-5, o-series, GPT Image), Gemini (Imagen 4, Veo, Nano Banana), DeepSeek, Kimi, Grok, and 12+ providers

Agentic CVE remediation platform for Node.js. Correlates threat intelligence, applies policy-governed fixes, and delivers auditable remediation outcomes across CI/CD pipelines, agent workflows, and service portfolios.

MCP server that detects cross-signal correlations in project health — the thing no individual check can tell you.

CodeSlick CLI tool for pre-commit security scanning — 308 checks across JS, TS, Python, Java, Go

MCP server for AI Sightline — monitor brand visibility across AI platforms (ChatGPT, Google AIO, Perplexity, Claude, Gemini, Copilot)

MCP server for React Native spatial integrity and layout audit. Provides Ground Truth for your component tree by detecting margin pollution, token deviations, bridge crossings, and render traps. 15 tools for structural analysis, layout auditing, performan

Unified MCP server for email access across Gmail, Outlook, iCloud, and IMAP — search, send, organize, and batch-manage emails from any AI assistant

Search bills, legislators, committees, and events across all 50 US states, DC, and Puerto Rico via MCP. STDIO or Streamable HTTP.

MCP server for web search, extraction, crawling, and research — powered by Novada's 100M+ proxy IPs across 195 countries. Works with Claude, Cursor, VS Code, and any MCP client.

VidLens MCP — video as a queryable local asset for AI agents. 45 tools across YouTube, social video URLs, local media, semantic search, transcription, and visual search.

Turn Claude research output into designed PDFs. Cover page, table of contents, code blocks that hold across pages. MCP server and Claude Code skill for markdown to PDF.

AI-powered idea & task manager for developers. Brainstorm → structured tasks → AI agent execution. CodeMirror editor with slash commands, MCP Server, Claude/Gemini/Codex support. Local-first SQLite, cross-PC Git sync.

dwic — the design auditor for Claude Code. MCP server that audits your design tokens + markup for WCAG contrast failures, mandated-accent drift, and structural gaps, then remembers findings across sessions.