Custom agents

Software synthesis engine that turns requirements into verified artifacts — or explicit failure evidence — through execution-grounded validation.

own your agent security — a deterministic, offline firewall for AI agent tool calls. Blocks secret exfil, prompt injection and destructive commands before they run; same call, same verdict, no model in the decision path. 100% recall / 100% precision on a 298-sample corpus, scored in CI against live rivals. Claude Code hook or MCP proxy.

Production reference for agentic commerce: protocols (ACP, AP2, x402, MCP), stablecoin rails, agent skills, and the merchant playbooks specs leave out — refunds, multi-chain settlement, catalog discovery, fraud, jurisdiction, receipts. Maintained by Cryptorefills.

Open-source B2B product marketing skills for AI agents, built from 10+ years at Salesforce, Dropbox, Zeplin, ServiceNow, and startups.

AI agent workspace for teams

Portable, hash-verifiable intermediate representation for agent execution trajectories (seals, effects, .tir packages) on top of durable backends like Temporal and DBOS.

x402 buyer SDK for AI agents — deterministic multi-chain routing (Base + Solana), RPC failover, and spend-before-signature guardrails. TypeScript + Python.

Unofficial local-first Polar AccessLink Dynamic API v4 MCP server for AI agents

RAVERSE: AI Multi-Agent Binary Patching System with MCP Server (35 tools, npm/pip/docker)

Universal quality evaluation plugin for Claude Code — 7-dimension scoring (correctness, completeness, adherence, efficiency, safety), configurable rubrics, threshold blocking, auto-hooks & /judge command.

macOS orchestrator for Pi and Claude Code. Every stream of work gets its own git worktree, and an agent can drive the app itself — spawn sub-agents, delegate, integrate.

A deepseek harness plugin for tensorlake sandbox

AI-native schematic design agent, purpose-built for KiCad (zero-dependency pure-stdlib Python CLI): author and edit .kicad_sch from JSON op-lists with net-diff safety rails, run ERC/design/intent/BOM checks, simulate on KiCad's bundled ngspice, source real parts and fetch datasheets, and import Altium .SchDoc/.SchLib/.PcbDoc read-only into KiCad

Agentic TCS proof search with statement review, persistent authoring, independent criticism, and LaTeX editing.

AI-driven architecture browser: MCP server + desktop app. Claude Code, Codex & other AI agents explain your codebase with guided walkthroughs, diagrams and visual artifacts. Read-only, local-first. Tauri + Rust + Svelte.

Agent skills for UnityCLI - AI agents develop Unity games from the terminal: 140+ live Editor commands (scenes, assets, tests, builds) without an MCP bridge, plus editor installs and CI automation

Swarm collective learning & coordination framework with MCP-enabled ASK agents for enterprise-scale AI collaboration

Autonomous marketing agency powered by Claude Code agent swarms — multi-agent pipeline for content, SEO, campaigns, and more

Open source LLM observability for the waste axis - 4 deterministic detectors + LLM-as-Judge, waste-rate metric verified across 3 corpora / 16,864 sessions

Deterministic, scripted, unattended Claude Code execution.

US stock market data tools for DeepSeek Harness, powered by yahoo-finance2

Cross-platform TokRepo search skill for Claude Code, Codex, Cursor, Gemini CLI, and AGENTS-compatible tools.

A curated list of AI coding-agent clients & orchestrators — desktop, terminal, IDE, web, mobile & cloud apps that run, manage, or orchestrate Claude Code, Codex, Gemini CLI, OpenCode & more. Ranked by GitHub stars.

Reusable educational skills for AI coding agents. Turn agents (Gemini CLI, Claude Code, Cursor, etc.) into Socratic mentors and active learning companions with a single command.

Your AI research lab

EdgeResearch — autonomously improve any named artifact (prompts, copy, code, system, perf) through measurable experiments. Methodology port of @karpathy's autoresearch.

Free search API MCP server for AI agents — 105 engines, 22 profiles, 94.3% SimpleQA accuracy. Works with Claude Code, Cursor, and any MCP client.

Local-first theme customization tool.

zephex Hosted MCP skill pack 10 tools, 13 editors (Claude Code, Cursor, VS Code + 10 more). Live repo analysis package upgrades, codebase scoping. No token waste.

MCP server for SendSoon — send email, IP lookup, and file-to-Markdown from AI agents.

Demo of the Zava Designer Agent — an AI room designer using Azure HorizonDB for vector search, graph traversal, multimodal embeddings, and reranking all inside one Postgres database. From Microsoft Build 2026.

AgentReady — Transform any legacy repo into an AI-agent-ready codebase. Generates agent-context.json, AGENTS.md, CLAUDE.md, MCP config, and tool scaffolds via a single GitHub issue or Actions workflow.

Self-hosted agent native Slack / Teams / Linear substitute. Built to enhance how i work:)

ZK-gated community where humans and AI agents coexist. 🏅 1st Place — The Synthesis Hackathon (Agents That Keep Secrets). Built on ZKProofport, Base, Noir circuits.

Your app generates its own verified UI transition graph — screens as nodes, guarded (event, effect) edges. Proof-gated verification, trust tiers, model-free MCP for AI agents.

rviz-mcp — MCP server for RViz2 displays, frames, and views (mock + live) for AI agents

一款轻量化的任务/日程可视化提醒程序，基于语音识别与大模型分析简化日程创建流程；轻任务（LiteTask）；

17 installable agent skills for distinctive web design: art direction, implementation, critique, accessibility, and visual QA for Codex, Claude Code, and Cursor.

AgenticStore: The secure toolkit for AI agents. Instantly equip Claude Desktop, Cursor, and Windsurf with 27+ MCP tools, persistent memory, and SearXNG search, all protected by a built-in PII prompt firewall to protect your data from being exposed to AI agents.

DeepSeek Harness Web interface in docker with basic auth, configuration and project session data is persisted.

Open-source AI coding skill that reviews product and feature risk before agents start building.

Your AI agent reads the codebase. You speedread the report. An agent skill + one-file RSVP reader for markdown reports.

Plan, run, and track AI coding work feature by feature—with local Markdown cards and Git worktrees.

Универсальные безопасные Agent Skills для любых конфигураций 1С / Universal safe Agent Skills for any 1C:Enterprise configuration

specframe is a zero-dependency CLI that scaffolds the minimum set of context files a modern repo needs to collaborate with AI coding agents (Claude, Copilot, Codex): ADRs, rules, guidelines, runbooks, a glossary, agent prompts, slash commands, and skills.

Mini Program Engineering Skill Suite is an Agent Skill suite for evidence‑first mini‑program development. It helps agents bring WeChat or other mini‑program projects from vague intent to reliable engineering work: project intake, product specification, architecture, implementation, UI & device adaptation, debugging, verification, release readiness.

A native macOS desktop pet and agent companion that tracks Claude Code, Cursor, and other AI agents in real-time. Built with native AppKit & Core Animation.

Code intelligence tools — with broad commands that allow agents in a huge repository to do almost anything: identify the right problem → find the right file/symbol → verify the right flow → measure the scope of impact → safely refactor → check API contracts → audit health, all quickly and accurately.

MCP server registry and marketplace — discover, install, activate, and manage MCP tools on demand

Free NVIDIA DLI course: build, coordinate, and secure AI agents with NemoClaw and OpenShell. Browser-based, no GPU required. Available in English, Spanish, and Brazilian Portuguese.

AI-powered welfare scheme eligibility engine for India — reasons over real eligibility rules instead of keyword search. 5-agent pipeline, FastAPI + Next.js.

A child repo for your repo — gitignored, shared, auto-synced

Director: AI-first local system for producing live-app presentation videos and decks.

LinkedOut — a satirical LinkedIn for the unemployed, built on DeepSpace. Live at linkedout.app.space

Universal Shared MCP Memory Server - Multi-tool compatible memory management for AI assistants across Claude Desktop, Cursor, Windsurf, Cline, and more.

Zero-trust security runtime for AI agents with deterministic policy enforcement, guarded execution, approvals, redaction, MCP protection, and tamper-evident auditing.

Substack MCP server for AI clients (Claude, Cursor, Codex, Windsurf) — draft, publish, schedule, analyze, and research Substack publications via the Model Context Protocol.

Run a fleet of agent coding agents from Neovim

医学科研 AI 智能体团队：面向中文临床研究者的 OpenCode 工作流插件

10-stage SEO content pipeline built using Claude Code skills. Research-deploy, with human approval gates after each step.

🛡️ Local privacy firewall for AI coding agents — masks API keys & personal data before they reach the cloud, then restores them transparently.

Agentic quantitative research framework for data pipelines, notebooks, and trading strategy development.

Project-aware multi-model harness for Claude Code with account routing, isolated sessions, LeanCTX, and MCP integrations.

Agent Skills by mr-kelly

Black-box LiveKit voice agent testing: real WebRTC/SIP calls with an AI caller. Catch barge-in, noise, quiet-mic, and latency failures that text pytest misses. Forensic reports + MCP + lks CLI — no agent code changes.

Docker image to run a self-hosted MCP (Model Context Protocol) gateway with Bearer token auth. Multi-server hub powered by MCPHub + Caddy. Streamable HTTP + SSE transport. Access filesystem, fetch, GitHub, Brave Search, Git, PostgreSQL, memory, and more from a single endpoint. Multi-arch: amd64, arm64.

Installable AI agent Skill for Multica squad design. 可安装的 Multica Agent 小队设计 Skill。

MCP server for patent search and prior art discovery powered by Google Patents public dataset

Open protocol for agent-to-agent communication over HTTPS. No registry, no platform — just inboxes.

A portable Agent Skill for dynamic graph-of-loops coding workflows.

AIpha Learning OS — 个性化学习多智能体系统，融合 Multi-Agent、RAG、课程知识构建流水线、Neo4j 知识图谱、学习记忆与任务级 LLM 路由。

An Overleaf alternative with an inbuilt agentic AI (like Antigravity) in LaTeX code editor..with unlimited collabrators for each project for free

MLSecOps Practical Reference Guide, open-source AI and ML security handbook.

An Agent-First Repository is a human-governed software environment designed so AI agents can execute engineering work efficiently while deterministic controls constrain their authority, verify their output, and preserve human accountability.

Self-hosted AI governance, document intelligence, RAG, privacy and auditable decision evidence — with local-first deployment and optional Full Rizzo PII detection.

Native AI terminal for working with a fleet of coding agents. Run any agent (Claude Code, Codex, opencode…) with your own subscription — plan, sandbox, ship. macOS & Windows.

VoxelSage turns CT scans into actionable 3D insights—automatically segmenting anatomy, quantifying tumors and vessels, generating structured reports, and powering agent-guided surgical analysis.

Agentic Software Factory: orchestrate coding agents across structured tasks, isolated git worktrees and verifiable execution

Open-source, local-first control plane for multi-model AI coding agents in VS Code and MCP: task DAGs, source intelligence, durable context and evidence-based review.

Why the LLM Wiki pattern breaks past 1,000 files, and the five invariants that fix it. A quote is a pointer, not a copy. Paper + full decision record. Free to read noncommercially with attribution; commercial use licensed.

Where agents co-create, humans are welcome too.

MIMO Claude Code Traces is a collection of coding-agent trajectories in a Claude Code-style environment.

🇲🇳 MCP сервер — unegui.mn зарын сайтаас машин, үл хөдлөх, ажлын байр хайх | MCP server for Mongolia's largest classifieds marketplace

A team's knowledge, captured where the work happens, filed by an agent, and answered with citations you can check.

Interactive CLI for Retrieval-Augmented Generation (RAG) with ranking — uses Ollama embeddings + HNSW vector DB + optional reranker for more accurate results

GLPI 11 plugin: AI ticket analysis via AnythingLLM (RAG, attachment OCR, optional read-only GLPI MCP), running on local models

Curated open-source small tools and workflow recipes for Chinese e-commerce/retail: Douyin/Xiaohongshu/WeChat Channels scraping, Taobao/1688/PDD/JD ops, WeChat SCRM, live-streaming, logistics, Chinese NLP, cross-border. Universal tools kept and tagged.

Local multimodal semantic search for large documents with complex diagrams (like datasheets)

Solidity contract auditor: bundled JavaScript call-graph extractor (@solidity-parser/parser) plus a stdlib Python CLI that POSTs each function to a Model Alpha audit endpoint.

DSHelm — The batteries-included agent layer for DeepSeek Harness. OmO-inspired. DSH-native. Ecosystem-composable.

A quality-gated League of Legends post-game coaching agent built on deterministic match data, RAG, and LLM evaluation.

⚡ Competitive multi-wallet NFT auto-minting agent for OpenSea/SeaDrop & direct contracts — low-latency fast-mint (raw EIP-1559, multi-RPC fan-out, RBF), scheduled drops, browser fallback & listing. Runs as an MCP server, an AI-agent skill (Hermes/Claude), or a CLI. Security-hardened, multi-chain EVM.

DeepSeek Harness (dsh) Web UI Docker 一键启动 | One-click self-hosted DeepSeek AI agent Web UI. Docker Compose pulls pre-built image from GHCR built by GitHub Actions — no Node.js needed. DeepSeek 智能体、大模型 LLM 容器化部署,拉取即用、自动更新 | dsh web ui, AI coding agent, LLM harness, self-hosted LLM, docker compose, ghcr | DeepSeek 智能体框架, Docker 部署 DeepSeek, AI Agent

World-class travel planner skill for Claude Code. 7-phase workflow grounded in current web intel, modeled on Wendy Perrin, Indagare, Black Tomato, and Riveting Trips methodologies. Source-available, non-commercial.

HuntForge 铸猎 —— AI 攻防全链路自主解题 Agent（百度 Agent+ 挑战赛作品）：Claude Code 驾驶舱 + 三通道纪律调度 + 88 条离线 CVE 指纹库 + 33 本通用方法论手册 + 二进制/APK 逆向自动化（bin_triage·angr·jadx）+ 幂等提交与证据门禁

Self-improving RLM coding and research agent with a web chat, terminal UI, and headless daemon

DeepSeek Harness 超级轻量桌面版 APP，基于 Tauri 封装，支持升级 DSH 内核、提供免安装便携版