MCP servers and skills: search, filter by where they run, and compare.
Playwright for coding agents. Test that your skills, MCP servers, and CLIs actually work when an agent uses them — sandboxed YAML suites, A/B experiments, CI gates.
PythonApache-2.0★ 154No key
Benchmark ClaudeCode/Codex token usage across compression scenarios
SkillApache-2.0★ 8No key
Scalable and extensible reinforcement learning for LM agents.
SkillApache-2.0★ 122No key
开源 Windows 本地 AI 工作台｜OpenAI-compatible + Claude CLI，99 个桌面/Office 工具，可回退文件操作与可视化多 Agent 工作流。
SkillApache-2.0★ 40No key
Your coding agent is smart. Your legacy codebase is the problem. Reify gives it the context to understand and safely change large, unfamiliar codebases — 100% locally.
SkillApache-2.0★ 3Few starsNo key
decision-engine, autonomous-agents, modernbert, pytorch, model-context-protocol, system-1, developer-tools, pr-gating
SkillApache-2.0★ 2Few starsNo key
Next-Generation Distributed Multi-Agent Runtime & Swarm Substrate uniting Super-Memory, Cyclical Swarms, 24/7 Durable Daemons, Multimodal Streaming, VLA Safety Governors, and Benchmark Evaluators.
PythonApache-2.0Few starsNo key
SkillApache-2.0★ 4Few starsNo key
Robot Framework library for testing MCP servers, Agent Skills, Hooks, SubAgents, and coding-agent CLIs — provider-agnostic via LiteLLM, BFCL-grade tool-call matching, statistical N≥10 by default.
PythonApache-2.0★ 1Few starsNo key
Make every coding agent faster, cheaper, and more accurate with a persistent local code graph that learns from each session.
GoApache-2.0★ 2Few starsNo key
🧪 the Evaluation Operating System — OEC (Observe → Evaluate → Control) for foundation models, agents, and business objectives
SkillApache-2.0★ 1Few starsNo key
Claude plans, Codex / OpenCode Go / DeepSeek execute: a quota-aware dispatcher that routes coding tasks to the cheapest capable agent, in isolated git worktrees with hard budgets.
SkillApache-2.0★ 2Few starsNo key
An open-source, cost-efficient (<zsh.50/task), and extensible Agentic Software Development Life Cycle (AIDLC) framework
SkillApache-2.0Few starsNo key
KV-faithful Dynamo agent trace replay and closed-loop coding-agent load generation
SkillApache-2.0Few starsNo key
Claude Code skill: route subagent workers by task profile. Frontier reasoning for planning, judging and synthesis; commodity pricing for execution. Based on Anthropic's benchmarked Fable-orchestrator/
SkillApache-2.0★ 1Few starsNo key
Optimize any AI agent’s skills, tools/MCP, and prompts against your own evals.
SkillApache-2.0★ 66No key
面向可信个人开发机的 MCP 编程运行时：Agent Transport、Windows Workbench、证据优先的项目连续性 | Personal-workstation MCP coding runtime
PythonApache-2.0★ 18No key
Evidence-first code intelligence for AI coding agents: verified claims with evidence, honest unknowns, a name check for Python and Java (invented APIs, Mixin targets), PDF/Office documents and screens
PythonApache-2.0★ 3Few starsNo key
Stops your coding agent searching from zero — a 22ms local file ranker. No API key, no index, no network call.
SkillApache-2.0★ 4Few starsNo key
Local search over your code and documents for AI agents. One binary that indexes your repositories, PDFs and Office files and serves them over MCP.
SkillApache-2.0★ 5No key
Verinoda + Claude Code, together: Verinoda with verinoda-live, a Claude Code mod that keeps the code index fresh, nudges the agent to use it, reviews commits, and shows it in a pane with an animated m
PythonApache-2.0★ 1Few starsNo key
sư dụng trên nền tảng aider phát triển và tối ưu
PythonApache-2.0★ 2Few starsNo key
Who stands behind it
