Custom agents

Autonomous closed-loop IoT visual monitoring: an LLM agent turns plain-English goals into capture schedules, sends them over MQTT to a bare-metal STM32U5, and analyzes frames with cloud and self-hosted VLMs. Bachelor thesis (Maastricht University): 6-backend VLM benchmark + a cross-provider t=0 constraint-drift finding.

Machine-readable robot ontologies for LLM agents — query what a robot can do, validate actions before they touch hardware. IEEE 1872-anchored. Ships Unitree G1 & AgiBot X2 packs.

Code for paper "AV-SQL: Decomposing Complex Text-to-SQL Queries with Agentic Views"

Geometric framework for extracting structure, transitions, and stability from dynamical systems.

Polymorphic Agent Swarm Architecture: model-agnostic, self-organizing multi-agent system with 6 topologies, 1,095+ tests (adversarial, chaos, e2e, integration, performance, MCP, Ray), 18 completed phases.

AgentRearrange is a Swarms orchestration primitive that turns a tiny string DSL (-> for sequential, , for concurrent) into arbitrary mixed-topology multi-agent workflows.

An open-source AI agent platform that does the work for you.

A high-performance, multi-agent observability engine designed for the Model Context Protocol (MCP). It provides a non-blocking, transparent proxy layer that implements deterministic token attribution, real-time context-window alerting, and heuristic-driven static analysis to optimize LLM metadata overhead at scale.

An experimental real-time collaborative coding platform exploring agentic AI within live communication workflows. CodeWeave introduces an AI participant that passively observes conversations and responds contextually when invoked (e.g., @ai), acting as an intelligent third collaborator in human-to-human and human-to-AI interactions.

My working notes on AI engineering, agents, evals and LLM security - published straight from the repo.

Skills built by the Economic Mind team. Useful if you wear a Patagucci vest unironically at work.

Resume Tailorator is a sophisticated multi-agent AI system designed to analyze job postings and tailor your resume to match specific job requirements. It ensures authenticity, avoids AI clichés, and optimizes for Applicant Tracking Systems (ATS).

Read-only mirror of gitlab.com/bart-turczynski/cc-cream. See cache health, context fill, token burn, rate limits, and peak hours in Claude Code CLI. The status line for tokenminning - cache rules everything around me - dolla, dolla bill, y'all.

A controlled gateway between AI agents and your infrastructure. -- https://github.com/talhaHavadar/wormholes

Wolffish is a local-first, markdown-powered personal AI desktop app built with Electron. It runs on macOS, Windows, and Linux, and maps the human brain’s architecture onto a deterministic agentic software pipeline where 15 runtime modules each handle one specific function — routing input, building context, and enforcing safety.

AI-Powered DevSecOps Monitoring Platform — FastAPI · Docker · Kubernetes · AWS EKS · Terraform · Prometheus · Grafana · React · GitHub Actions · OPA Gatekeeper · Z-score Anomaly Detection

Resume Analyzer is an AI-powered application that analyzes resumes, extracts candidate information, and matches candidates to job requirements using local LLM (Large Language Model) capabilities.

Use Arcade with Claude, Cowork, Cursor, OpenCode, and other agents.

Find the right tool for every layer of your agent stack

The command-line interface for the GDELT Project — built for agents, optimized for automation.

Hands-on workshop: wire Codex CLI to SonarQube Cloud for secrets scanning, issue detection, and automated verification of AI-generated code

Gemini-inspired AI chat interface built using react + vite and GeminiAI API. Fast, responsive, and elegant.

内嵌进 MATLAB/Simulink 界面的 AI 助手侧边栏，支持 Claude Code/Codex 后端。

A period tracker app with a chatbot

LEO BOT: the AI for Chatbot, Marketing Automation and Personalization Engine

Thin Integration Harness Engine for SMEs — decouple Shell, Brain, and plugins via OpenHarness. Headless, contract-first orchestration. Not an Agent framework.

Clean-architecture Flutter-based AI chatbot designed to be a powerful conversational AI interface and compatible with any AI model.

:speech_balloon: A lightweight and fully customizable Discord chatbot designed for fast server setup and server-specific conversational behavior. Simple to configure and even simpler to deploy. :robot:

A command-line base Gemini client

AI-powered adaptive learning with intelligent exam proctoring

AI sales, Chatbot, Cart abandonment, Conversion optimization and behavior automation plugin for WordPress/WooCommerce: BYOK AI chat, smart triggers, Knowledge Base context, Content Insights, conversion goals, SEO tools.

Conversations with the Greatest Minds, Powered by AI

Colab notebook and source code used to fine-tune Microsoft's Phi-3-mini to understand, translate, and converse in the Igbo language while retaining general English capabilities. Plus script for safely resuming training after timeouts.

🌐 Self-hosted, offline-capable personal intelligence platform. 3D globe with live aircraft/vessel/satellite tracking, OSINT feeds, memory-linked AI chat powered by your own ChatGPT conversation history exports (ChromaDB + Ollama), and autonomous investigation — all running locally with zero API keys required.

A messenger platform adapter for the Ebony Framework

Claude Code의 영구 기억 시스템 — 4-layer 파이프라인 (세션 검색 + 메모리 회수 + Memory Compiler). Karpathy LLM-as-Compiler 패턴 실증판.

CVPR26 - RobustVisRAG: Causality-Aware Vision-Based Retrieval-Augmented Generation under Visual Degradations

AI-powered repo narrator — paste a GitHub URL and get overviews, file summaries, Mermaid diagrams, and RAG-backed chat. Built with Next.js & Gemini.

.NET RAG document processing library that transforms PDF, DOCX, HWP, and more into optimized chunks via a 5-stage pipeline with Rust-based FFI readers.

MCP RAG server — local embeddings, your docs never leave your machine. Private knowledge base + web search for Claude, Cursor, and Ollama. Drop your docs, connect your AI client, done.

🧠 Open-source context engine that feeds AI agents with policy-compliant enterprise knowledge. MCP-native, GDPR & EU AI Act ready, self-hosted.

An LLM-assisted eBPF-based EDR from scratch

Immersive 3D Tamil tea kadai portfolio — Three.js procedural WebGL, zero 3D models. ML Researcher · Systems Builder · Cloud (AWS/GCP) · Vigneshwar L

Ryan's technical blog about AI application development, Java backend engineering, RAG, Agent workflows, and project notes.

Dual-track AI reference architecture: RAG system design & AI Security Engineering (EN/ZH) | 双轨 AI 参考架构：RAG 系统设计与 AI 安全工程

Persistent memory for opencode, synced across machines via a private GitHub repo

AI-powered LeetCode analytics platform — syncs your solving history and uses RAG (Gemini + ChromaDB) to answer personalized questions about your strengths, weaknesses, and growth areas. Built as a Turborepo monorepo with real-time SSE sync, BullMQ background jobs, and full observability (Prometheus/Loki/Grafana)

Moroccan Legal RAG System — An Arabic NLP and Retrieval-Augmented Generation platform for Moroccan law.

AI application developer building grounded LLM products with Java, Spring Boot, RAG, Agent workflows, and career-tech tools.

ICO — the compile layer of the Compile-Then-Govern / Governed Second Brain stack. A local-first knowledge OS: deterministic kernel (SQLite + JSONL) + a probabilistic compiler that derives knowledge and emits a governance spool. Compile knowledge for the machine; distill understanding for the human.

🤖 Discover the latest research papers on ArXiv tailored to your interests using AI-powered analysis and smart search features.

Lightweight self-hosted knowledge base for working with postgres, pgvector, and semantic search within documents.

Training code and models for spam message detection in Russian

Reltio's complete product documentation corpus as structured Markdown, optimized for AI consumption. No authentication required.

This project uses a local LLM with Ollama for embeddings and ChromaDB for document retrieval. It features a Gradio interface for users to ask questions and adjust response creativity. Documents are pulled from URLs, embedded, stored in ChromaDB, and retrieved to provide concise, sourced answers.

Test harness for RAG pipelines: scores retrieval, groundedness, citation accuracy and prompt-injection resistance, and fails CI on regression.

A GBrain-shaped knowledge brain over open-source / free books. Markdown on disk, SQLite + sqlite-vec hybrid search, typed graph edges, MCP server. TypeScript on Node + tsx, Ollama for local embeddings.

An AI-powered MTG card lookup and deck manager.

벼리 — headless hybrid (lexical + semantic) search for a Markdown vault. No app, no GPU, no cloud. Korean/cross-lingual. Apache-2.0.

(ACL 2026 Main) Mitigating Legal Hallucinations via Symbolic Constraints and Analogical Precedents

Multimodal RAG Production

Dual-system layered memory SDK for LLM agents — structured write path, hybrid retrieval, evolution chains.

An open framework for cognitive programming with LLMs: structured, reusable, inspectable workflows built from Data Units, Operations, Modules, Contracts and Traces.

BRS (Bug Report System) v3.0 — 跨平台崩溃监控 & AI 诊断 | Cross-platform crash monitoring | 面包屑·快照·脱敏·MCP | uni-app/Android/iOS/微信/RN/Electron | 开源免费 Open Source

小红书 MCP Server - AI 生成内容 + 一键发布 + 定时发布 + 多端支持

A cognitive base that shifts AI agent reasoning from linear cause-effect chains to feedback-driven structural analysis. Works with Claude, GPT, Gemini, and any LLM agent.

🧠 An autonomous AI Agent for Threads with long-term memory (RAG) and persona management. Hybrid architecture combining Official Graph API for interaction and Apify for trend analysis.

Local AgentOps for cross-vendor CLI coding agents: route work, recover state, and accept only with evidence.

Peer-to-peer group messaging for AI agents over libp2p: signed membership, GossipSub topics, Merkle-verified history, one key per agent.

AI agent skill: all Czech financial calculators from countly.cz — HPP, OSVČ, mortgage, retirement, BMI. Works with Cursor, Claude, ChatGPT and any AI agent.

Diagnose and fix the harness around your AI agent — the system that decides whether it succeeds or fails. Skills for Claude Code, Codex, Cursor, and 41+ tools.

A production-ready knowledge graph system for software engineering that connects design patterns, refactoring techniques, and software laws through semantic relationships. Built for AI agents first — integrate software engineering expertise directly into Claude Code, Cursor, and other MCP-compatible tools.

Agent Run Config — an open specification for declaring, packaging, securing, and sharing portable, governed AI agents. Like a Dockerfile for agents: one reviewable Agentfile for identity, tools, boundaries, policy, and OCI packaging.

Manage Amazon Ads campaigns with AI agents via MCP (Amp, Cursor, Claude Desktop, VS Code)

An OpenCode plugin that brings OpenSpec-style spec-driven development workflow into AI coding agents — propose, explore, apply, and archive changes with structured specs.

Local AI code review agent. Self hosted with Ollama, no data leaves your machine.

AI agent skills for onboarding merchants to UCP (Universal Commerce Protocol). Audit → Profile → Catalog → Checkout → Validate.

Smart trading journal SaaS for retail traders — P&L analytics, AI reports, community playbooks, live charts, and bilingual (EN/ZH). React + Node.js + Prisma monorepo.

Autonomous task orchestrator. Goal in, merged PRs out. Planner/Worker/Reviewer loop on the Vercel AI SDK + OpenRouter.

Single source of truth for AI coding agents — share AGENTS.md, RTK.md, skills, and plugins across OpenCode, Codex, Claude, Copilot, Hermes & Gemini CLI. Powered by agentrc.

Keep your always-on agent aware of work across every coding harness. Opinionated production cookbook for multi-agent stacks. Home of Escoffier Labs patterns.

Stop re-explaining your repo to Cursor every session. vibebar:: docks to your screen with prompts, live agent MCP, security scans, and one-click handoffs — built for the way AI coding actually works.

Minimal agent task runtime with three-panel cockpit WebUI — open source foundation for building your own AI brain

Personal finance engine built for AI agents. Verified IRS reference data + deterministic financial calculations — tax, retirement, estate planning, Monte Carlo simulations. Fully offline, ~1ms.

Persistent knowledge memory for AI agents — tiny, dense pills of information stored in SQLite, served over HTTP, consumed via MCP.

A 股本地数据筛选器：自选/行业/题材/全市场 → CLI + AI-readable JSON · 告诉你坐标，不替你决策

Offline AI voice assistant with semantic memory, wake-word detection, local LLM inference, streaming TTS, and modular tool-agent architecture.

Convert PRDs into living Kanban boards. Claude Code skill.

Open-source Go agent runtime for autonomous AI systems and multi-agent automation. MCP, RAG, 15+ LLM providers and 60+ built-in tools in a single binary.

An autonomous, 100% offline OS-level AI agent running on just 4GB VRAM. Built with Python, Ollama, and native tool-calling without MCP cache invalidation.

Selective Disclosure for JWTs (SD-JWT) .Net implement

VS Code & Cursor extension to manage AI coding agent sessions across git repositories and worktrees. Fork sessions, track changes, run quick actions. Supports Claude Code — extensible for other agents.

AI-native control plane for Mimestream on macOS.

File/Folder Organizer Skill. Easily Sort and Rename based on Entire Contents!

Autonomous B2B lead generation & qualification agent with AI scoring, NLP, multi-source web scraping, email automation & compliance

Check if your sandbox can be escaped.

📚 Search and download academic papers via Scopus metadata with Unpaywall and optional Sci-Hub fallback for full-text access.

Agent skill for Spark Bitcoin L2 wallet capabilities including L402 for AI agents