MCP servers and skills: search, filter by where they run, and compare.
Cross-provider LLM consultation MCP server (OpenAI + Anthropic + Gemini + xAI Grok + Jan local). First concrete impl of Trellis DSM-26 AngularTechnique substrate.
PythonMITFew starsNo key
Andrej Karpathy's LLM Council as a portable agent skill for Claude Code, Cursor, Copilot & co — built entirely from his repo, posts and research
SkillMITFew starsNo key
A Claude Code skill that fans one hard question out to several frontier models (GPT-5.5, Gemini, Kimi, GLM) blind and in parallel, then has Opus 4.8 judge and synthesize one grounded answer. Multi-mod
SkillLink onlyFew stars
Multi-model panel and judge (OpenRouter-Fusion style) for Claude Code, opencode, and any MCP host. Diverse models answer in parallel, then a judge reconciles them into one better answer.
PythonMIT★ 6No key
PythonMIT★ 4Few starsNo key
Compose agents like plain values. A TypeScript toolkit that wires LLM calls, tools, and functions into typed steps over eight providers. No framework, no ambient state.
Node.jsApache-2.0★ 4Few starsNo key
🧠 Graph Memory — MCP Knowledge Graph Service for AI agents. Graph-first approach with Neo4j + Qdrant RAG, ontology-driven extraction, multi-tenant. Built by Cloud Temple.
SkillApache-2.0★ 7No key
Early warning system for heat and cold risk. Predicts personalized thermal mortality risk by age, sex, comorbidities and activity, with an ML ensemble (XGBoost + LSTM) calibrated via conformal predict
PythonApache-2.0★ 5No key
A blind panel of models + a judge, for decisions where being wrong is expensive. Verifies before it synthesizes; counts consensus by model family, not by head.
SkillMIT★ 3Few starsNo key
Harness-agnostic, evidence-gated agent orchestra for autonomous ticket delivery — renders into Claude Code & OpenAI Codex from one neutral core.
SkillMIT★ 2Few starsNo key
Python package that makes it easy to spin up a custom Retrieval-Augmented Generation (RAG) pipeline.
PythonGPL-3.0★ 4Few starsNo key
PythonMIT★ 1Few starsNo key
Budget-native multi-model coding agent with live steering, local Ollama, and trusted MCP tools.
PythonApache-2.0★ 1Few starsNo key
The AI agent firewall: One face watches what enters, one watches what leaves. Defends against prompt injections, exfiltration, and scams with a multi-layer ensemble.
SkillMIT★ 2Few starsNo key
Claude Code plugin: MCP server that offloads bounded LLM tasks to cheaper local (LM Studio, Ollama, vLLM, llama.cpp) or remote (OpenRouter) models with profile-based configuration and ensemble mode
SkillMIT★ 1Few starsNo key
Loop Engineering for Kaggle, built entirely on the Claude Code ecosystem (Skills + Hooks): scout → survey → hypothesize → experiment (Colab) → submit. Exploratory, science-backed, human picks the comp
SkillMIT★ 7No key
SkillMIT★ 3Few starsNo key
Unified Context-to-Execution pipeline: Obsidian semantic search + GitNexus code impact analysis + Agent Skill Bus orchestration
SkillMIT★ 5No key
White-label marketing strategy brain — ensemble AI, calibrated market signals, margin-aware budgets, self-contained briefs for execution agents. Every decision audited.
MCPLink only★ 1Few stars
Ensemble de serveurs MCP (Légifrance, Judilibre, image, système, Gris) utilisables automatiquement par Demeter (v2.4 minimum) et n'importe quel LLM acceptant les serveurs MCP
SkillAGPL-3.0★ 1Few starsNo key
Multi-agent document and form extraction system with parallel LLM-as-a-judge consensus, adversarial forensic auditing, and interactive UI dashboards.
SkillApache-2.0Few starsNo key
Who stands behind it
