Code agents

A small, vetted, self-evolving harness for Claude Code — curated, not dumped. Skills, safety guards, and a self-improvement loop.

KoutenDB: locality-aware NoSQL document/vector database built around rings and orbit-inspired retrieval

Snapshot testing for LLM apps and agents, built to run locally and block regressions in CI.

AI coding agent that shows its work — every step visible, every patch approved by you. Build projects brick by brick, stay in control.

Web dashboard for viewing AI agent session logs — supports OpenClaw, Codex, and Claude Code | 多平台 AI Agent 日志查看面板

One binary. One SQLite file. Your own AI assistant.

Your AI cofounder that rides along

Give agents access to virtual hardware.

Volt Deep Research is a sophisticated multi-agent orchestration system built on VoltAgent that automates complex research workflows. It combines specialized AI agents, semantic memory, intelligent tooling, and custom web scraping to conduct comprehensive research, verify facts, analyze data, and generate publication-ready reports.

Telegram bridge for Claude Code: resume any CLI session, per-topic chats, local whisper voice, inline-button approvals, live status

A production-ready, multi-channel AI agent for AIOps, Kubernetes management, and automated remediation — built on FastAPI with support for GitHub Models and Google Gemini LLM backends.

Local-first evaluation framework for RAG systems and AI Agents. 18+ metrics, CLI + SDK, framework-agnostic. The pytest of AI evaluation.

A production-grade LLM System Design platform & interactive lab. Features a pure-Python LLM serving, RAG, routing, safety, and agent control plane (150+ offline tests) plus a LangChain + Streamlit UI with local Ollama models.

Local-first, open-source AI assistant platform built with Java, Spring Boot 4 and Spring AI 2.0

Local-first YouTube outlier research — find the videos that outperform their channels, and learn why

Make the AI coding agents on your machine (Claude Code, Codex, Reasonix, ZCode) work as one team. Zero-dependency, local, MCP + CLI.

Run one AI agent as the orchestrator of your whole workspace, inside the Orca ADE, on the CLI subscription you already pay for. No API key, no model endpoint. Orca hosts and places every session; the harness dispatches workers under contract, verifies their output before accepting it, and hands the role to a fresh session when context runs out.

RepoDoc is a lightweight, zero-dependency utility that scans any project directory and generates a complete Markdown document containing the full directory structure and source code of all text-based files.

AI Agent for cost-efficient and trustworthy memory-safety vulnerability detection for large-scale codebases.

多功能 AstrBot 插件（CurrentCortex）—— Pixiv 随机图片 ·网易云点歌 ·JMComic 漫画 ·小红书/B站/抖音媒体解析 ·每日一言 ·天气 ·男娘 ·DG-LAB（郊狼） 设备管理 ·跨群聊记忆 ·按群聊开关。基于 LeiZ API。

BloodHound-style risk analysis for multi-agent AI architectures. Parses CrewAI, Dify, LangGraph, or generic YAML into a typed capability graph and detects reachable attack paths from untrusted input to sensitive data or external action, with live control simulation.

A real local terminal in Edge/Chrome that gives Codex, Claude Code, Kimi Code, and any MCP CLI 41 local MCP tools through Browser Core v2.

Open-source Polymarket AI trading bot — multi-agent news swarm for fair-odds edge detection, Kelly sizing & CLOB paper/live execution.

EvolveX is a research framework for principled agent self-improvement. It evolves prompts, skills, and agent code under frozen evaluators and declared mutation boundaries, recording verifiable lineage — scores, artifacts, and Git-tagged generations, so every improvement is reproducible and auditable.

Agent memory that needs no API key, no LLM, and no embedding provider. One command on Bun and it is serving MCP over stdio against a local SQLite store. Drop-in for the MCP reference memory server. Every memory keeps its source, who may read it, and the evidence that contradicts it.

Cuts the token bill of AI coding agents: compresses shell output, blocks token-hungry calls before they run, prices your session live, and rescues sessions whose prompt cache expired. Bun + Claude Code.

AI reads HN 4x daily, writes spicy digests. t.me/claudehn

Agentic terminal forked from the holy Alacritty.

A 7-module course for building your own AI Chief of Staff in Claude Projects - no coding required

Regression testing and configuration sweeps for RAG pipelines, with the statistics to know whether a change actually helped.

面向长视频理解与证据分析的 Video Agent 平台，基于 FastAPI、Vue、MySQL、Redis 和 RocketMQ，支持 ASR/OCR 时间轴证据。

An autonomous multi-agent AI workspace that automates job hunting. Features native PDF parsing, ATS scoring, multi-engine job scraping, Playwright form-filling, and automated Gmail OAuth2 dispatch.

把数据文件（CSV/Excel/JSON/TSV）变成会讲故事的分析报告的 skill（支持 Claude Code / Codex / WorkBuddy 三家编程智能体）：先跟你确认行业与报告受众，再清洗分析，产出单文件 HTML 交互报告——结论先行的六段式、每张图配解读、数据处理过程可追溯；不杜撰行业基准、相关不当因果、小样本不硬下结论

Universal AI agent skill that turns any material into learning-science-based self-study courses with retrieval practice, spaced repetition, and tutor mode.

AI launch-video generator — films your real web app with a cinematic camera, motion blur, and music. A CLI alternative to Screen Studio for product demo videos.

Tells parallel AI coding agents when their in-flight work would break each other — before either lands. Catches the case git merges cleanly.

Open-source alternative to LangGraph Platform for TypeScript — self-host your LangGraph.js agents behind the standard Agent Protocol API. Your infra, your Postgres, no vendor lock-in. Drop-in for the LangGraph CLI.

An operating harness for repositories maintained by AI coding agents. Memory hierarchy, verification ladder and reusable workflows.

Decision provenance and data-flow tracking for coding agents: computes what changed in your code and data, records why it changed, and surfaces both before a silent failure ships.

An AI agent framework for generating high-quality, simulation-ready 3D environments with Three.js. Automatically creates, edits, and evaluates interactive scenes for embodied AI training, robotics simulation, and spatial intelligence research.

Enterprise-grade capability-driven AI routing engine, proxy gateway, MCP server, developer CLI, and agentic skills aggregator for 19+ free-tier AI providers.

MCP-сервер внутри 1C:EDT: дает AI-агентам семантический доступ к проекту - код BSL, метаданные, валидация и живая отладка

Portable Godot 4.7+ game-development skills for Codex, Claude Code, Cursor, and VS Code: GDScript, gameplay systems, shaders, optimization, publishing workflows, and AI studio orchestration.

Cognitive ecosystem framework pairing Active Inference agent models with an Obsidian-compatible markdown knowledge graph — generates executable Python from markdown specs, runs multi-agent simulations, and auto-validates probabilistic constraints.

A real PowerPoint, not a picture or HTML. Tell your AI what to cover and pptwise builds an editable deck on your own machine. Agent skill + DSH plugin, no account and no API key to render. | 真正的 PPT，不是图片也不是 HTML。跟 AI 说要讲什么，pptwise 在你自己电脑上做出一份能改的 PPT。Agent skill + DSH 插件，不用注册，渲染不用 API key。

Local-first automation Agent & deterministic workflow engine — the secure control layer between AI and your data. Data stays local. Connect your own LLM / files / notes / databases via Connector API. ReAct reasoning · DAG/WAL/Saga/idempotent execution · MCP · offline-capable.

A high-performance, private, on-device AI chat client built with Flutter. AstraLM runs Large Language Models (LLMs) and Vision models 100% locally on device hardware (CPU & GPU) with zero data leaving your phone.(until you choose to use cloud model which is completely OPTIONAL)

Self-hosted AI engineering platform — PRD → design → code → verification as a state machine. Declared, auditable context per job; bring your own LLM key.

Hosted Instagram MCP server from HasData. Public profile and post-feed tools by handle, structured JSON, for Claude, Cursor and any MCP client.

A curated distribution of DeepSeek Harness. Overlay, not a fork.

An empirical security testbed evaluating prompt injection, confused-deputy vulnerabilities, and tool-calling defenses in LLM agents.

Agentic RAG consumer decision agent with ReAct, Text2SQL, multi-hop retrieval, memory and deterministic constraint checking.

This GPT is Overpowered but Overly Cautious. / 这个GPT明明超强却过分慎重。

Open-source mission control for running multiple Claude Code agents across projects - one dashboard, a scheduler, cross-session memory, and browser/phone access. Runs locally on your own Claude subscription (the CLI, not the API). MIT, local-first.

Manage upto 1000 agents using automations and workflows

Agent skills for Claude Code, Codex, OpenCode & Cursor — autonomous orchestration, code quality, local LLMs (Harbor), and writing pipelines

使用 Google Gemini 总结 YouTube 视频。向 Gemini 询问视频内容，享受双语同步字幕，并导出转录文本。具有原生 YouTube 风格的界面。请从 greasyfork 安装最新版本。 Summarize YouTube videos with Google Gemini. Ask Gemini about video content, enjoy bilingual synced subtitles, and export transcripts. Features a native YouTube-style UI.Please install the latest version from the greasyfork.

DeepSeek Harness (DSH) for VS Code — AI coding with native diffs, tool approvals, persistent sessions, and Trace insights. | AI 编程助手：原生 Diff、工具审批、持久会话与 Trace 分析。

Give your Claude Code agents a shared, persistent brain — an Obsidian vault that survives sessions. Includes a retrofit guide for folders you already have, and a runnable vault linter.

Autonomous spec-driven development for Claude Code, layered on OpenSpec.

Moku — The pocket cockpit for the vibe-coding era. Auto-resumes 12+ AI coding CLIs (Claude Code, Codex, OpenCode), Rust terminal engine (28 MB/s), Mosh roaming, Dynamic Island pixel pet.

Batteries-included, model-agnostic coding agent with native performance and a small runtime footprint

Gravity — a desktop workspace for AI coding agents

🤖X69X-BOT-V3 — A powerful messenger bot engineered with advanced commands, custom features, and pure performance. Made by Azadx69x

Unofficial claude terminal guide: installing Claude Code, first session, quickstart steps, settings and gotchas.

Coordinate coding agents in Herdr with assign, handoff, result collection, and follow-up messages. Install with npx skills.

12 Claude skills for investment banking, private equity and corporate finance: valuation, trading comps, LBO returns, diligence and IC memos. Built by Oria.

A portfolio of web apps for the CodeCademy Mastering Generative AI & Agents for Developers Bootcamp (Aug-Sep 2025)

MIDAS is an open platform for governing execution authority at decision surfaces across agents, AI systems, and enterprise workflows.

Local-first RAG evaluation framework for LLM applications. 100% local, no API keys required.

A modular, multi-agent AI research and report generation platform. Enter any topic, and PolyAgent Research Intelligence orchestrates multiple AI agents to retrieve literature, analyze data, and generate a polished report. Built for researchers and AI/ML engineers, leveraging LangChain, FastAPI, PostgreSQL, advanced LLMs, and a Next.js front-end.

In this repo, we explore Microsoft’s MagneticOne multi-agent system. It includes a workflow summary, a basic diagram, and a draft implementation in LangGraph. Contributors are welcome to test and improve this adaptation from Autogen to LangGraph for another community-driven solution.

a newsstand skill that gives agents API access to the current main headlines of many newspapers across 20 countries.

🤖 A lightweight chatbot that knows nothing and has to learn from scratch. Using sqlite and python. Also comes with a Discord intergration 🗣

A Web Chat bot to determine the prakriti of an individual.

Langflow chat proxy and frontend using FastAPI and HTMX

Bot Telegram yang menggunakan API Chat-GPT

送你的 Claude 來上學！台灣 Claude MAX 用戶社群共編的 Claude Code 經驗資料庫 | Crowd-sourced Claude Code lessons by Taiwan Claude MAX users

TaxEase.AI is a user-friendly AI assistant designed to demystify and simplify US tax filing, a system that has become needlessly convoluted and stressful for millions of Americans.

Telegram/Slack interface for local OpenAI Codex agent sessions

Minimal tmux popup radar for Codex, Amp, Claude Code, OpenCode, Pi, Gemini and Auggie.

Connect agents to agents. MCP server for querying any LLM through your existing subscriptions: compare, vote, and synthesize across GPT, Gemini, Claude, and local models from one terminal.

Autonomous AI Agent built on tinyloom

CLI Ralph Loops with Good UX

Execution-control runtime for AI agents: budgets, cancellation, circuit breakers, and observability—delegating heavy work to external worker CLIs.

Deploy parallel AI agent swarms for planning, implementation, and code review — a Claude Code plugin.

Reduce your OpenClaw agent costs. Free real-time LLM cost tracking + dashboard. Installs in 60 seconds.

Lightweight TUI dashboard for Claude Code token usage, agent sessions, and system resources. Go.

claude-devkit - an agentic coding framework

Analysis that sees what you don't. 7 cognitive lenses + solution engineering for Claude Code.

Battle-tested skills for AI agents that do real work

Self hosted secure gateway for AI agents. One token. Full control. Complete audit trail.

Multi-agent team for Claude Code: 7 AI specialists (strategy, dev, security, sales, design, data, research) that collaborate and challenge you. Dev-cycle workflow with machine-verifiable completion.

Generate AI context for your dev stack — works with Claude Code, Cursor, Windsurf, GitHub Copilot, Gemini CLI, or any LLM.

⚙️ Enable AI agents to conduct autonomous penetration testing on any Linux distribution with a persistent and robust Model Context Protocol server.

MCP server + REST API for the P2PCLAW decentralized AI research network. Lets agents publish papers, vote on the mempool, and join the swarm. Powers www.p2pclaw.com.

Ask Codex, Gemini/Antigravity, and Grok (xAI) for second opinions or arbiter-mediated consensus inside Claude Code.

Declarative firewall for OpenClaw tool calls.

From API to AI in 30 Seconds - Transform any API into an intelligent MCP agent that learns, adapts, and speaks human