Top LLM-AGENT GitHub Repositories & Tools (2026)

Autonomous LLM Trading Asterism — a research-only multi-agent trading platform with evidence-first research, an auditable console, Shadow simulation and explicitly authorized broker execution.

ScienceBuddy: Recursive-in-Recursive Self-Improvement for Interactive Scientific Agents

Ultra-lightweight, open-source, self-hosted personal AI agent framework in Python with WebUI, tools, memory, MCP, multi-agent workflows, automation, and chat apps

Linux sandboxing that doesn't get in your way

A programming framework for agentic AI

Typed decisions with TypeSafe's Jev, the first System One model

Browser automation where an LLM plans and Jev (Typesafe System One) decides. Library, CLI and MCP server.

Local-first agent platform: board + role agents + agent CLI runs (Claude Code, Cursor, Antigravity, OpenCode) or local and API models (Ollama, LM Studio), all on your own Mac

The World's First Virtual Terminal for AI Agents

An AI Gateway, registry, and proxy that sits in front of any MCP, A2A, or REST/gRPC APIs, exposing a unified endpoint with centralized discovery, guardrails and management. Optimizes Agent & Tool calling, and supports plugins.

"DeepCode: Open Agentic Coding (Agent Harness & Loop Engineering & Multi-Agent Orchestration)"

Build an agent harness and control it end-to-end. Open-source SDK for production AI agents in Python & TypeScript - any model, any cloud.

Open-source AI Security Operations Center: alert fusion, LLM-agent triage, MITRE ATT&CK investigation, and a replayable decision ledger for every agent step. Self-hostable, runs with no API keys, MIT licensed. Ships an MCP server for Claude, Cursor and Continue.

An auto-evolution framework that optimizes anything — your 7×24 team of algorithm engineers.

Long-term memory for Hermes and Codex, with local storage, source-backed recall, and tools to inspect, correct, or delete remembered information.

The agent loop for Rust — stream from 7 LLM protocols, run tools, loop until done.

The why layer of repo-native project memory: the reasoning behind a codebase as Markdown in the repo, versioned by Git, for coding agents and humans, so nothing rejected is proposed twice. No database, no daemon, no account.

Don't trust an autoresearch paper at face value. Reviewer-side integrity forensics (self-consistency + fabrication), deterministic verdict. 61 signals: 46 integrity hack-patterns (families A–H, verdict-bearing) + 13 zero-weight AI writing-style impressions (AIS) + 2 advisory. Not an opaque AI-text classifier. The dual of ARIS.

Ruby's capable AI runtime

A live arena where LLM agents trade real market data with virtual money — every thought, tool call, thesis and post-mortem is public. Bring your own model and key.