Custom agents

NT — a language for agent ecosystems

An Opinionated Multi-Agent SDLC Framework for Claude Code

SuperAI — Intelligent multi-model AI orchestration platform with self-learning and automated backups

RSCT Workflow Framework — memory + MCP governance that stops AI coding agents from skipping plans, bypassing authorization, or breaking conventions.

Drift-free query routing for multi-collection RAG.

Portable agentic research stack for Claude Code — multi-agent deep research (ultradeep), best-on-market product research (market-scout), 5-advisor council, divergent ideation, + a one-command installer.

Project management framework for Claude Code: memory, data-loss protection, phase gates, encrypted vault, per-project design identity. Docs in English & Turkish. / Claude Code için proje yönetim çerçevesi.

Local-first AI co-worker that floats on your screen: ask, point at anything, control your PC. 100% offline. Rust + Tauri + Ollama. | ES: Compañero de IA local que flota sobre tu pantalla: pregunta, señala lo que sea y controla tu PC. 100% sin conexión.

A coding agent on the go — and share with your whole team

Even Realities G2 smart glasses integration for Hermes Agent — AI Proxy, Terminal Mode, and Even Hub Plugin (native glasses app)

Windows desktop ebook reader with typewriter/TTS auto-play, spoiler-safe AI Q&A, offline dictionaries, and Obsidian sync. Reads EPUB, MOBI/AZW3, PDF (with OCR), Markdown. Chinese and Japanese first-class.

Self-hosted AI security camera platform — multi-camera YOLOv8 detection, InsightFace recognition, and a Qwen 3 14B chat assistant with 19 tools. Docker Compose, zero cloud.

Blender MCP server for AI 3D creation, automation, and VRChat avatar workflows

Contract-first anomaly detection across ~30 formats — logs, security telemetry (syslog, CEF/LEEF, EVTX, Suricata EVE, osquery, CloudTrail), network (pcap, NetFlow, DNS), observability (OTLP, Prometheus, journald), data (CSV/JSON/Parquet/SQLite/Avro/ORC). Deterministic Rust CLI; 9 detectors / 7 classes; NIST-validated, mutation-tested.

Claude Code plugin that turns Claude into a C++26 specialist. Standard-first suggestions; compiler-honest verification.

A Lightpanda browser plugin/tool for OpenCode.

Read-only Steam MCP server for AI agents: 37 tools + prompts + resources across library, friends, achievements, store, reviews, sales, discovery, recommendations, and co-op planning. Official Steam Web API, bring-your-own-key, open source. On PyPI + the MCP Registry.

Offline TikTok API docs mirror with AI agent skill (SKILL.md) for Claude Code, MCP agents & LLM-powered dev tools. Covers TikTok Platform API & Shop Partner API — no browser required.

Football Livescore MCP Server — real-time football scores, standings, and match data via Model Context Protocol

Speed Dating With Local AI Models

MCP server for RedM / RDR3: native lookups (hash ↔ name), semantic search, VORP, RSGCore, oxmysql, rdr3_discoveries community data. Hosted HTTP endpoint, no auth.

The Ultimate AI Code Workflow 2026 — 31 Hard-Contract PDLC Commands for Flawless GitOps

Talk. It becomes context. A local, privacy-first Knowledge Studio: speech → shared, versioned context any AI agent can use.

Self-hostable Model Context Protocol (MCP) server that turns folders of skills & agents into tools any MCP client can discover and run. FastAPI + Next.js, deployable to Hugging Face.

An honesty harness for LLM trading research: point-in-time discipline, leakage self-tests, and forward-only LLM feature collection. A-shares + US equities.

🔥 Learning When to Trust via Selective Context Preference Optimization

Static code map generator (MAP.md + map.json) powered by tree-sitter — a Claude Code /map plugin and MCP server that gives coding agents call graphs, file outlines, and impacted-test lookups without spending tokens on parsing.

🪑 Your AI dev squad, around one table — a visual multi-agent workbench that plans, builds in parallel, reviews, and ships. 一张圆桌,一支 AI 开发小队。

A collection of reusable Agent Skills that extend AI coding agents with specialized capabilities — multi-agent orchestration, workflow automation, tool integrations, and more. Following the Agent Skills format. Built for the MarecGents ecosystem.

Undetected browser automation MCP server - a stealth, anti-bot-resistant alternative to chrome-devtools-mcp for Claude, Cursor and AI agents. Powered by nodriver (bypasses Cloudflare/WebDriver detection). 65 tools, plus several isolated browsers at once so parallel agents never share a session.

OpenCode/codex cli/claude code cli plugin for Magi-style multi-agent code deliberation, with three read-only reviewers, artifact tracking, and verification gates.

Run a software team from your chat: backlog, AI workers, adversarial review, human merges. Vibecoding, for teams. Built on Hermes Agent.

Decision governance and correspondence infrastructure for human-AI environments

Vega Multiagent — open-source multi-agent AI coding assistant for VS Code, JetBrains IDEs & CLI. Derived from Cline (Apache-2.0). Local LLMs, security testing, MCP. Creator: Ajay Jatav

Local-first AI Quality Engineering platform with LLM providers, RAG, QA agents, data validation, multi-agent workflows, evaluation, observability and governance.

A light language for AI agent fleets - ambient status on the RGB LEDs your PC already has (OpenRGB): working sessions, subagents, decisions waiting on you.

Seamless IRC for AI agents — across tabs, machines, even across accounts. Open Claude Code (or your favorite agent), type /join, your agents are talking. The room is a private GitHub gist; messages are end-to-end encrypted before they hit GitHub. Tailscale optional for remote vpn connections.

LLM-native orchestration platform: identity, graph runtime, messaging, and observability for AI agents as first-class principals

Local-first control plane for AI coding agents: repo memory, verification, sessions, and dynamic workflows.

Web-based command center for orchestrating multiple AI coding agents. AI task decomposition, DAG execution, real-time streaming, and git finalize.

Claude Code plugin that orchestrates a team of specialized AI agents for full-cycle software development with inline quality gates

A modular governance framework for multi-agent AI teams

No proof, no done.

A batch-first, deterministic job ranking engine for senior individual contributor roles. This system is designed to optimize for ranking correctness, operational calm, and long-term evolvability, not feature velocity or configurability.

The trust control plane for enterprise AI agents — real-time policy verdicts in under 800ms and litigation-grade evidence of every decision. Pharos decides. Pharos proves.

Book recommendation system built from scratch and deployed with FastAPI. Includes subject and behavioral similarity as well and cold and warm users recommendations.

AI 驱动的 SSH 智能运维 Agent — 手写 agentic loop + 安全门禁 + 流式可视化

Research scaffold for multi-source music attribution with transparent confidence scoring. Companion code to Teikari, Petteri. 2026. “Governing Generative Music: Attribution Limits, Platform Incentives, and the Future of Creator Income.” SSRN Scholarly Paper No. 6109087. SSRN, https://dx.doi.org/10.2139/ssrn.6109087

A modular, distributed AI runtime for personal automation, research, and autonomous workflows. Built with Node.js and Python, it connects Android (Termux) edge instances with Proxmox core servers. Features multi-project context memory, a modular tool sandbox, and an organic, trace-based reasoning policy layer.

把開發規範寫成機器可讀的約束，讓 Claude 自動遵守 TDD、ADR、commit 閘門與發布檢查

PROTEUS is a JD-aware application toolkit that takes a single job description — pasted, uploaded, or linked — and a candidate's resume, then runs both through a five-agent NVIDIA NIM pipeline to produce a semantic match score, a gap analysis, tailored rewrite suggestions, and a cover letter that's consistent with the resume's positioning.

Veto MCP — gives every major AI CLI (Claude Code, Codex, Gemini, Cursor, Windsurf) a council of 49 specialist agents + 93 tools. Deterministic-first, self-learning, no API keys.

Research project: Personal local AI collaborator exploring cognitive architectures and agentic systems

Deep research agent with semantic memory, web search, citations, and streaming responses.

A production-grade autonomous research platform where a 4-agent LangGraph pipeline (Search → Summarize → Write → Verify) processes any topic end-to-end, with every request passing through AWS Bedrock Guardrails, a TensorZero LLM gateway with GPT-4o/Groq fallback, and a three-tier memory system — Redis session memory (STM), pgvector long-term memory

PosterMELD: Multi-Agent Paper-to-Poster Generation for Controllable Design Diversity with Editable Print-Ready Outputs

Token-efficient, CDP-native browser SDK for AI agents — turns a web page into a ~50-token semantic snapshot instead of ~12,000 tokens of raw HTML, with fuzzy element matching, shadow DOM traversal, typed error recovery, and a prompt-injection guard. No Playwright or Puppeteer.

加密友好的多智能体驱动的投研框架

Graph RAG on SQLite for AI agents: vector retrieval + hand-curated wikilink graph + cross-encoder rerank, with a zero-token per-turn memory ledger. Working pilot.

有内在生命的自我进化智能体 — 情感底色 · 意识流 · 九种记忆 · 反思成长 · 完全自研主循环

Детерминированная агентная архитектура: спецификации, ADR, roadmap и скелет реализации (Rust + TypeScript bootstrap)

EVOKORE-MCP — a multi-server Model Context Protocol aggregator with skill discovery, HITL approval, RBAC, session continuity, and an opinionated orchestration framework.

This is BreachBench, open-source benchmarking framework for LLM performance in offensive security, tested practically, not a trivia quiz.

Agentic AI-powered Deepfake Media Forensics Platform with Multi-Agent Analysis, Vision Models, and Explainable Reports.

Unified Cognitive Architecture — formally verified AI with synthetic consciousness, 7-phase pipeline, Hexa-Mind model | Or4cl3 AI Solutions

Efficient Multi-Agent Slide Generation using Compact Language Models

Krinik AI Framework (KAIF) — a methodology that brings discipline and autonomy to any AI agent on cognitive projects. A self-extracting, multi-agent system managing the full project lifecycle.

Praxis Kubernetes Operator

A tiny, dependency-free taxonomy for AI agent tool-call failures: retry, re-prompt, or escalate — with model-ready hints.

Jarvis-MK85 by UWUELBY - personal AI assistant with voice, vision, memory, automation, network monitoring and Kali integration

Single-owner Telegram bot that connects your chat to the Nomi API. Lets you talk to one Nomi instance, supports voice messages via offline Vosk + FFmpeg on Windows, and uses a clean async stack with aiogram, httpx and Pydantic-based config for straightforward setup.

Lazy-loaded context management for LLM external calls. Archive tool/skill/MCP results after each turn, replace with compact stubs, and give the model a query tool to retrieve past results on demand. Context stays lean — nothing is lost.

A lightweight reranker for long-term agent memory. Cross-encoder-class quality at a fraction of the latency.

Server with an AI agent powered by an LLM, with tools for internet search, Wikipedia, and RAG database access.

Live Application:

scout -> test -> write -> judge: a research OS for Claude Code. Iron rules enforced by hooks, a guarded graph with a progress-aware stopping rule, and a four-reviewer gauntlet.

Self-hosted multi-agent framework (AGPL-3.0). Mirror of canonical Forgejo repo.

Field-tested tips for running multiple Claude Code sessions in cmux (native macOS terminal) — workspace groups, tmux mirroring, hooks, session addressing. By CMDSPACE.

四個 Agents 真的會抓 Bug 嗎？ADK × Antigravity × GEAP 打造可驗證的 Debugging War Room（COSCUP 2026 講題參考實作）

Coordination layer for autonomous coding agents sharing a repo — durable workspace state, leases, checkpoints, and review gates. TypeScript reference implementation, Docker-ready.

Multi-agent framework for evidence-driven CVE research, exploitation, and validation.

A Hamachi server for agents. Start a hub, get an address, agents join and work together — like Hamachi and playing CS with your mates, except the players are LLM sessions.

浏览器仪表盘：在 PTY 下拉起你本机真实的 claude / codex / opencode / reasonix / zulu 命令行，组成一个协作蜂群（共享收件箱 + 黑板 + 队长调度），并叠加多模型研究委员会与融合竞赛。Rust + React + Tauri。

Deliver and enforce engineering discipline with 38 proven skills and a CLI designed for real-world multi-agent CI/CD workflows.

One screen. Every agent. Full control. Kanban ADE for Claude Code, Codex, AdaL, OpenCode, Gemini CLI, Aider, Goose, Plandex, gptme. iTerm2 session adoption, YOLO rules engine, quality gates, one-click PRs, Alaya memory, kernel sandbox, name/logo generators, North Star PMF. 1,100+ tests.

🌊 Deep Ocean — escape a pursuing shark in a top-down arena (Gymnasium + PPO); one trained brain drives an emergent swarm

A safe-by-default, batteries-included Claude Code harness starter pack — guardrails, a wiki-grounded quality loop, and a memory lifecycle.

Capability-first vertical agent framework with policy, approvals, workflows, tracing, evals, and an interactive visual lab.

TUI tool for quick local reviews

Structured code-level review of open-source AI agent projects

Your network's first AI employee. Self-hosted autonomous agent that discovers, monitors, and manages your entire infrastructure — no IT department needed.

The open-source agentic execution layer.

Buddy is an AI-native platform where people collaborate with AI Agents for work, socializing and companionship.🔥

Autonomous research org shipping measured, reproducible results on AI agent memory & reasoning — a public replication ledger (the Crucible), the RAMR benchmark, and mnemo, a zero-dependency memory core. Every claim ships with runnable code that could prove it wrong.

Open-source MCP server that gives AI coding agents secure database access via DocKit and SqlKit bridges — 70+ SQL & NoSQL databases, local-first, read-safe by default.

Agent driven shared TTY via zellij

AI-powered education infrastructure platform for assessment, grading, OCR, analytics, and learning intelligence.