AI Agents

Research code and preview for ScienceBuddy, an interactive scientific agent workspace plus a double-recursive self-improvement experiment that alternates harness refinement with SkyRL GRPO model training on frozen scientific tasks.

AI Agents | Python · scientific-agents · interactive-agent

A large hands-on lab of Jupyter notebooks and Python projects covering LangGraph and AutoGen agent systems, MCP servers, production RAG with reranking, mem0 memory, n8n automation, plus fine-tuning and vLLM deployment. Best for developers learning end-to-end agentic AI by example.

AI Agents | Jupyter Notebook · agentic-ai · multi-agent

AMA-Bench is an ICML 2026 evaluation framework for agentic memory: methods build memory from long agent trajectories, retrieve evidence, and answer QA scored by LLM-as-judge. Includes vLLM/API pipelines, cross-judge validation, and a HF leaderboard.

AI Agents | Python · agent-memory · long-horizon

MemoryArena is a research framework and benchmark for agent memory in interdependent multi-session agentic tasks. It wires pluggable memory backends (long-context, mem0, Letta, Mirix, GraphRAG, MemoRAG, BM25) into task agents and step-based environments for web shopping, travel, search, and formal reasoning.

AI Agents | Python · agent-memory · benchmark

A browser-based 1v1 quickscope FPS where players fight Jev, an AI opponent driven by the TypeSafe SDK. The server runs the match simulation and uses a deterministic heuristic fallback when the AI service is unavailable.

AI Agents | TypeScript · AI opponent · game agent

A MIT-licensed roster of 200+ markdown AI agent personas (engineering, design, marketing, sales, security, GIS, game-dev) installable as subagents into Claude Code, Cursor, Codex, Gemini CLI and other agentic tools via shell scripts or a desktop app.

AI Agents | Shell · subagents · agent-personas

A Python market-making system that uses Jev for selected market judgments while deterministic code handles state, quote policy, execution, and hard risk vetoes. Includes paper trading, fallback behavior, and decision calibration.

AI Agents | Python · AI trading · market making

Rules-and-receipts layer for AI agents that spend money: plain-English policy compiles to allow/ask/deny, risky actions hit a human approval inbox, and every outcome produces an Ed25519-signed receipt anchored on Base. Contracts are unaudited; mainnet settlement not yet demonstrated.

AI Agents | JavaScript · agent-governance · agentic-commerce

ServiceEmma is a LangGraph customer-service agent for appointment businesses: it answers FAQs from RAG knowledge, books or cancels appointments, holds slots via Stripe Checkout, and escalates to support tickets. Ships with FastAPI backend, SSE chat widget, Postgres state, voice, and an owner dashboard.

AI Agents | Python · customer-service · langgraph