Code agents

The living wiki that writes itself.

Multimodal emotion assistant with visual emotion analysis, Qwen realtime voice interaction, and local RAG knowledge base.

Starter project for creating an API endpoint that performs web searches, scrapes results, and generates answers using LLM.

RAG with LLM agents for SQL & graph databases

Memory module for LLMs and Agents with MCP

Autonomous multi-agent framework for self-correcting image synthesis using 'Think → Generate → Critique → Refine' cycles with Gemini 3.1 and SDXL.

A Retrieval-Augmented Generation (RAG) system for automating Security Operations Center (SOC) log analysis. This project combines NLP techniques with vector search to process security logs, enabling semantic querying and visualization through an interactive Streamlit interface.

CodeMap is a CLI tool that generates optimized markdown docs and streamline Git workflows.

Hardware-Accelerated Cryptographic RAG System — GPU simulation of SHA-256 hash engine with AES-256-GCM encryption and Merkle tree integrity. NOT real ASIC hardware.

A RAG system for Contract Q&A that enables chatting with a contract and asking questions about the contract. It has an interface build with React and FastAPI in backend integrating rag-pipeline with Autogen agents and websockets for communication. Evaluation of the RAG is done using RAGAS.

A comprehensive Go library for managing multiple LLM (Large Language Model) providers with intelligent API key management, rotation strategies, and unified interfaces. GoLLMKit simplifies working with OpenAI, Anthropic, Google Gemini, and other LLM providers while providing enterprise-grade features like cost optimization, health monitoring etc

My personal notes, code and projects of the Udacity Generative AI Nanodegree.

A library to calculate similarity scores between two collections of text sequences encoded using transformer models for bitext mining, dense retrieval, retrieval-based classification, and retrieval-augmented generation (RAG).

AI-Tutor : LLM and RAG-Enhanced AI Tutoring for Various Courses

Versatile LLM Tool Use (Function Calling) package for Laravel, compatible with all LLMs, enabling LLM to execute actual code functions (unlike LLMs' built-in capabilities).

A Retrieval-Augmented Generation (RAG) system using Vector and Graph Database

RAG-Enhanced-NCERT-Tutor is an AI-powered tutor for NCERT curriculum, using Retrieval-Augmented Generation (RAG) to provide accurate and contextually relevant answers. It integrates retrieval techniques with a language model to enhance learning by offering detailed explanations and resources based on real-time queries from NCERT materials.

C# SDK for the Jina API -- embeddings, reranking, and search

Enhance SQL query generation using OCI Gen AI and Marqo. This project integrates with Oracle Autonomous Data Warehouse, leveraging DDL and comments for context-aware natural language queries, indexing and searching vectors, and executing optimized SQL commands.

Microservice chatbot LLM RAG

Building a RAG System connected to Confluence using Airflow, FastAPI and Qdrant

Investigating the vulnerability of Large Language Models (LLMs) to misinformation in Retrieval-Augmented Generation (RAG) systems by poisoning vector databases and analyzing LLM responses to identify potential weaknesses and exploitation risks.

Official dataset and code for "SyllabusQA: A Course Logistics Question Answering Dataset" (ACL 2024)

Production-grade LinkedIn CLI for humans and LLM agents

GNME (pronounced Gen-Me) is a lightweight framework that replaces static, task-based specifications with living Module Genomes. It is designed specifically for developers using AI assistants (Cursor, Claude) to build complex, evolving systems.

Browser-based Claude-Code-style agent harness for local uncensored Ollama models. MoE-fast, security-audit aware, MIT.

Spora is the lightest OSS AI Gateway. This product is managed by the startup Fisarum.

The self-healing AI scraping agent

Production-ready template for AI agent workflows with Hermes Agent, Temporal durable orchestration, and human-in-the-loop review. Ships with an exam grading demo.

Pure-Python toolkit for inspecting, editing, converting, and analyzing 3D print files, supports 3MF, STL, OBJ, PLY, and GLB. Detects source software (Bambu, Orca, Prusa, Cura, Fusion 360, etc.), checks compatibility across 15 CAD and slicer programs, and works with any AI coding agent (Claude, Cursor, ChatGPT, Manus). Zero dependencies beyond numpy

Local Sandbox Runtime for AI Agents. Run AI-generated code safely with OS, Wasm, and microVM isolation.

Secure MySQL MCP Server with Clean Architecture. Enables AI assistants (Cursor, Copilot, Cline) to safely interact with MySQL/MariaDB databases, featuring RBAC, audit logging, and intelligent query validation.

MSA Project

Vision-powered UX simulation engine for Claude Code. Renders pages in a real browser, captures 36+ screenshots across viewports, clicks through interactive elements, maps CTA funnels, tests signup flows, and scores across 7 UX dimensions. Replaces manual user testing with automated multi-viewport analysis.

⚔️ Disciplined orchestration plugin for OpenCode — automatic 4-tier pipeline, ClarityGate, dual review, runtime fallback

AI research-to-manuscript workflow for academic writing: literature search, source verification, research wiki, argument trees, outlines, drafts, reviews, citations, DOCX export, and thesis-destroyer workflow automation.

A plug-and-play framework for AI-assisted software development, enhancing context-aware collaboration in complex codebases. Perfect for tools like GitHub Copilot. 🐙✨

Claude Skill for querying and exporting WeChat 4.x chat history — DB schema, SQL templates, and LLM export format built on top of ylytdeng/wechat-decrypt.

🤖 Engage with an interactive Twitter AI agent to post tweets, retrieve recent updates, and schedule tasks effortlessly.

Universal security scanning skill for AI agents - finds hardcoded secrets, API keys, and vulnerabilities in any codebase. 44 patterns validated against GitLeaks, OWASP Top 10 mapping, Markdown/SARIF/JSON reports. Works across Claude Code, Cursor, Windsurf, and any agentic platform.

🧬 基于 AI 的 SBTI 毒舌人格测试：15 维度深度行为建模，支持人格碰撞分析与 AI 自定义诊断报告。比 MBTI 更接地气的互联网原生性格工具。原作者：B站@蛆肉儿串儿

Translate and adapt Superpowers for Chinese AI coding tools, with 20 skills and 6 China-specific workflows for better planning, coding, and debugging

业务Skill全生命周期管理工具——通过深度需求分析创建、修改、测试业务Skill。基于 Anthropic skill-creator 扩展，支持领域知识积累、三维度需求深挖、接口收集、Mock测试、迭代改进。

ov — Agent-first CLI for Obsidian vaults. JSON-only output, schema introspection, --dry-run safety. Built for AI agents.

A small toy skill for Codex and other code agents that adds a persistent ASCII companion pet to terminal coding sessions.

Autonomous orchestrator for Claude/Gemini/Codex CLIs. Markdown-driven queue with multi-provider fallback, policy gating, and full audit trail.

Live AI judge agent — watches hackathon demos in real-time via Gemini Live API, delivers British-accented commentary via TTS, scores with multi-model ensemble (Gemini + Claude + Groq), detects prompt injection attempts, and runs cross-team deliberation. Built for NEBULA:FOG 2026.

A governed coding agent CLI built on Claude Code. Runs inside an isolated Docker container with operator-controlled policy baked into the image, per-tool-call audit logging, and QA gates enforced before every session is accepted.

Control Mautic from the command line or any AI coding agent.

Practical AI recipes built on Eden AI runnable notebooks, 200+ models, swappable providers

AI agent for answering SQL-based questions using LLaMA and FastAPI

Teachable AI agent for controlled delegation, workflow learning, and operator-grade automation.

Block AI agent access to sensitive macOS paths and log all actions to protect private data during command execution.

AI Agent | AI Coding Agent | Agent Swarm | AI-Native Engineering Agent for Enterprise-Scale Codebase

Build a multimodal AI assistant for smart glasses using Whisper, vision-language models, and LLM hand on code with Google Colab, tested with Meta Ray-Ban Wayfarer.

Browser-first markdown plan review. Start local web review in one command, then let your agent continue through MCP.

Agent Skill for automatic CLI tool management - installs missing tools, audits environments | Claude Code compatible

AI 编程助手 Skills 学习教程 - 系统性学习和掌握如何创建、使用和分享 Claude Code Skills

AI Local Interactive Cross-device Engine. Same AI session, anywhere. Terminal ↔ Feishu for CLI agents (OpenCode/DeepSeek V4, Codex, Claude, Gemini, Kimi).

An experiment, not a tool: an autonomous agent on a local LLM that proposes changes to its own constitution and values; a human approves every one. Value layer (constitution, identity, skills, rules) sits behind a human approval gate, every change replayable. Runs on Ollama, holds up on a 16 GB Mac, no cloud LLM. Security by absence.

Give AI your browser. Check email, track orders, download receipts -- all authenticated, no API keys. Open source MCP server + CLI.

Connect Claude Code to messaging platforms (Telegram, Discord, Slack, LINE, WhatsApp) for bidirectional, remote interaction with your local AI agent.

⚖️ Fast CLI for searching Portuguese court jurisprudence (DGSI) and legislation (Diário da República) – AI agent skill for Claude Code, Claude.ai, Codex, and any LLM agent. Parallel search, full text retrieval, markdown/JSON output.

Multi-agent harness for long-running AI application development. Planner-Generator-Evaluator architecture with progressive simplification. Claude Code plugin.

Model Context Protocol (MCP) server for Google Workspace. Secure CRUD for Drive, Gmail, Calendar, Sheets, Docs, Tasks and People via AI agents.

Hermes Agent SaaS — the most complete open-source implementation of Nous Research's hermes-agent runtime. Multi-tenant, with shared and private Hermes Agent provisioning, 14 tools, vibe coding, mobile super-app. Built on hermes-agent + Kimi K2 Thinking.

AI hiring copilot built on Google ADK v2. Recruiter mode scores fit + drafts outreach. Candidate mode detects agency postings + generates full interview prep.

One-command auto-deploy of a Telegram AI agent (OpenClaw + Qwen via OpenRouter) on a fresh VPS. Handles proxy-per-key rotation and real-time rate-limit monitoring.

Java-native runtime for long-running AI agents with live execution control, human approvals, and audit trails.

微信OpenClaw一键部署 🤖 WeChat AI bot bridge for Codex device-code login, ACP, and OpenAI | 微信 AI 机器人桥接，重点支持 Codex 扫码式设备码授权登录

Connect NotebookLM research with Claude to generate structured content from URLs, PDFs, and trending topics for multi-platform publishing.

Stop prompting. Start engineering. Persona-driven installer that turns Claude Code into an engineering-governed development environment with enforced quality standards, specialist agents, and hooks.

Claude Code 源码资源合集 - 150+ 高质量资源 curated list

The first longitudinal benchmark for AI memory systems. Test LLM agents, RAG pipelines, MCP servers, and memory backends across 1,000 sessions.

Automated security auditing CLI for AI agent code — quarantine-first workflow for repos, packages, and agent tooling

Complete agent stack for Jaz. Plugins, Skills, MCP tools, and CLI for Claude Code, Cowork, Codex, Copilot, Cursor, and more.

Open-source AI security scanner for AI agents and skill packages. Detect prompt injection, data exfiltration, hardcoded secrets, and malicious logic before running AI tools. Supports OpenAI, Anthropic, Gemini, OpenRouter, and local Ollama models.

Niyam — AI-powered Applicant Tracking System (ATS), a modern alternative to Greenhouse, Lever, and SmartRecruiters.

Production-ready autonomous coding agent with React UI. Two-agent pattern, JWT auth, loop detection, checkpoints & rollback. Built on Claude Agent SDK.

AI-powered study tool for Claude Certified Architect exam. The codebase IS the curriculum.

A Git-backed, schema-validated handoff toolkit that gives each AI agent role (BA, SA, Dev, QA) a structured way to produce, validate, and pass SDLC artifacts to the next role.

AI-driven mobile app testing via MCP. Any AI coding agent can test any Android app - no test framework, no boilerplate, no testIDs

Forensic CLI for AI-era supply-chain attacks: XZ tarball diffs, prompt injection scans, MCP/CLAUDE.md/Cursor config audits. SARIF for GitHub Code Scanning.

Community-driven repository of reusable OpenClaw agents and sub-agents with validation workflows, fixtures, and tests.

Collection of subagents for Claude Code with configs spanning ideation, editing, version control

Co-Op is local-first business management software for company planning, research, files, customer work, and operational decisions. The installed desktop app owns the business workspace. The cloud backend owns account identity and license entitlement.

n8n 1.123.5 + Python 3.14 fully portable Windows. No Docker, no install. Extract → run start-n8n.bat → Python Code node works instantly. Offline/USB ready.

One command to make any repo agent-ready. Scaffolds .agent/ workspace, platform instruction files (Claude, Cursor, Windsurf, Copilot, Gemini), reusable skills, and MCP server for AI coding assistants.

Demos showing how to build with Feldera

Computer Experience Layer — MCP server for native computer use on macOS. Structured perception, reliable actions, continuous awareness.

Agent Reliability Engineering: applying SRE principles to AI agent systems. Evals, imp@k metrics, self-improvement, config versioning, transfer experiments.

100 game design lenses applied to your design artifacts via Claude Code or Codex. Auto-picks 3-5 relevant lenses, hedged multi-direction output. / 100 个游戏设计透镜审视设计 artifact。

Multi-agent AI Scrum team for Claude Code — native Mac app + CLI. Autonomous Sprints; you stay Product Owner.

An interactive analytics dashboard designed to visualize the difficulty distribution, vulnerability coverage, and structural diversity of the Xbow benchmark suite.

AI-powered research collaboration platform using Google Gemini. Find papers, generate proposals, and connect researchers in minutes.

Automatically detect and disable irrelevant Claude Code skills per project to save tokens and streamline your tech workflow.

Codex skill for agent-ready Linear project planning, execution ledgers, and issue-state hygiene

OpenDream — open-source memory consolidation for AI agents. Reads past any LLMs sessions, dreams across them to extract patterns, and writes consolidated memory into AGENTS.md so the next session starts smarter. Local SQLite, BYO LLM, no SaaS.

Teams-first multi-agent orchestration for Microsoft Copilot CLI

Repo-local verification protocol for AI coding agents: acceptance criteria, separate verifier roles, proof artifacts, and evidence-backed done claims.