Explore
MCP servers and skills: search, filter by where they run, and compare.
MCP servers and skills: search, filter by where they run, and compare.
Multi-agent deep-search skill (vocabulary-first → taxonomy → parallel expand → converge; hard length budget; layered deliverable) + multi-harness bench (Claude Code / Kimi / Grok vs single-shot Perple
SkillMITFew starsNo key
CLI, MCP server, and API-compatible interface for Perplexity AI
PythonMIT★ 20No key
Own your LLM's web search: a local search->fetch->rank pipeline that replaces hosted web-search tools. Measured: matches hosted accuracy at 66% lower cost and up to 88% fewer tokens, plus a precision-
PythonMIT★ 60No key
TouchDesigner Documentation MCP Server v2.6.1 - FIXED Python API tools! Features 629 operators + 14 tutorials + 69 Python API classes with working get_python_api & search_python_api tools. Zero-config
Node.jsMIT★ 80No key
Automatically pick an appropriate model and effort for every Claude Code subagent. Subagents inherit your session's model by default — a file search runs on the same premium model as your conversation
SkillMITFew starsNo key
A Claude Code skill that researches any topic using parallel web search agents and produces a sourced playbook
SkillMIT★ 5No key
The AI agent that flies on its own. Two commands to fly. 🪶
PythonMIT★ 2Few starsNo key
The Agent is an enterprise-grade Google Ads management agent running on Claude Opus
SkillMIT★ 31No key
The open-source reading OS that keeps everything you read as markdown files on your machine — with a frontier-model research assistant, your own replayable agents, and encrypted bring-your-own-cloud s
PythonAGPL-3.0★ 2Few starsNo key
A fast local vector store and code graph for AI agents — index your files once, keep it current as you save, and answer questions across all of it in milliseconds without leaving the machine.
SkillApache-2.0★ 4Few starsNo key
VibeCoding for 2 Years now, this is my AI/Skills/Promts/SubAgents/Agent-Orchestration/AgentWorkflows and MultiModel Speaking Agent TTS/SST . There are crazy good skills out there, but building them af
SkillMITFew starsNo key
Pure-Rust async runtime for LLM agents. Multi-provider (Anthropic/OpenAI/Ollama), tool-capable end-to-end, structured permissions, real MCP support, USD cost tracking, Files API, optional FS+search to
SkillMIT★ 19No key
Decision-model agentic search: fetch and parse deterministically, send only the decisions to a calibrated typed-decision model (Jev, or Laya locally). ~100-1000x cheaper than browsing agents. Part of
SkillMIT★ 2Few starsNo key
Multi-agent LLM pipeline that automates the AI-TPM job search end-to-end — built entirely with Claude Code orchestrating 11 specialist AI subagent
SkillMIT★ 1Few starsNo key
Lazy-loading skill discovery for Claude Code via MCP - mirrors Anthropic's Tool Search Tool pattern for Skills. 100% local Apple Silicon stack (MLX) or cross-platform Docker.
PythonMIT★ 6No key
Agentic AI copilot that finds freelance clients & drafts tailored proposals (RAG over portfolio, multi-agent, MCP) — human-approved before any send
SkillMIT★ 1Few starsNo key
Bulk YouTube transcripts → AI agent swarm. Ask any YouTube channel a question: one AI agent reads every video transcript in parallel and answers with timestamped citations. Rust CLI + MCP server for C
SkillMITFew starsNo key
Unified research-data acquisition MCP — search & fetch datasets across Zenodo, DataCite, NCBI omics (GEO/SRA/BioProject), and literature (PubMed/OpenAIRE) behind one normalized model.
PythonMIT★ 4Few starsNo key
a modular ai cli that unifies multiple model providers with built-in rag, embeddings, pluggable tools, and mcp support.
Node.jsMITFew starsNo key
An open-source retrieval engine implementing FABLE — structured, hierarchy-aware RAG
SkillApache-2.0★ 8No key
Cost-aware model routing for Claude Code: classify each subtask and route cheap work (search, boilerplate, docs, renames) to Sonnet/Haiku, reserve Opus for design, hard debugging, and final verificati
SkillMITFew starsNo key
How fast can a coding agent p-hack? Measured: on known-zero data, realistic searches manufacture p<.05 on 33-97% of null draws, in ~1 second each. Instrumented specification-search engine + 11 Claude
SkillMIT★ 8No key
SmartChain — multi-provider LLM conversation agent for Home Assistant (GigaChat, YandexGPT, OpenAI) with tool calling and streaming
SkillMIT★ 2Few starsNo key
Who stands behind it