Custom agents

A local board for Claude Code: Claude records tasks and asks its questions and reviews on cards, you answer them together, and work with acceptance criteria is not marked done until you approve. Sub-agent runs, side threads and Codex, OpenCode or Claude lane work show live on their cards.

Automated HackerOne hacktivity monitor - reports,playbooks, payloads, and telegram alerts for disclosures.

Hire specialists. Keep the final say.

A self-hostable workspace: pages, databases, issues and an infinite canvas over one set of records. Canvas 2D renderer, Yjs CRDT, local-first. AGPL-3.0.

A lightweight browser mission control for one Hermes agent: chat, sessions, cron jobs, skills, voice and read-only missions.

Your coding agents, working in a live pixel-art office: Claude Code, Codex, Command Code and more, local or in sandboxes. npx theagentoffice

stAirCase - The enforcement gate between AI agent plans and your codebase: approvals bound to exact bytes, a signed audit trail

The agent-agnostic AI tutor skill that writes into your Obsidian vault — works with Claude Code, Codex, Cursor, Windsurf, pi, and 70+ agents.

PostgreSQL/MySQL wire-protocol proxy for database guardrails for AI agents: stops destructive SQL before it reaches your database, no code changes.

Convert Microsoft Access databases to modern web applications — extract the intent, not the code. AI-powered pipeline: tables, queries, forms, reports, VBA modules → PostgreSQL + browser UI with structured intents for autonomous agent operation.

Self-hosted long-term memory for AI agents. Remembers conversations, tracks changing facts, and recalls context across sessions. MCP, Claude Code & Codex.

Portable token-reduction skill for any coding agent (Claude Code, Codex, Cursor, Aider, Claw, Hermes, Goose, Continue, OpenHands, vibe, Copilot CLI). Distilled from chopratejas/headroom. Drop-in agent rules + pure-Python helpers. 60-95% fewer tokens, no proxy required.

The open-source core of Hapimatic: reach and resume your Claude Code, Codex, Gemini, or OpenCode sessions from any browser or PWA, served from your own machine. | Solid PD | Philadelphia, PA

canli-mcp: 315 free finance tools for Claude and any MCP client, behind three (SEC filings, insider trades, 13F and 13D/G holders, options, FOMC, macro data, quant analytics, backtest validation). MIT, no API key, runs locally. Plus Canli Capital's open quant research, where every number links to its source file.

Free Cursor rules + AGENTS.md that stop AI (Cursor, Claude Code, ChatGPT) from writing broken Unity C#: missing usings, invented APIs, Unity 6 vs 2022 mixups. MIT.

AI has opinions. Humans have proof. An agent skill that split tests every material website change on real visitors instead of shipping guesses.

Let AI agents use an Android phone or iPhone: read the screen, tap, type, scroll and take screenshots. MCP server + CLI, with consent on the phone.

An evidence-based index of AI agent harnesses mapped by control function and autonomy.

A deterministic control plane for closed-loop agent workflows: routing, retries, approvals, and stop conditions in a loop file you can test.

Turn product details into sharp hooks, demo scripts, and sales copy. An open-source AI skill for affiliates, sellers, and creators—with 80 workflows, 24 hook formulas, and reusable project context. Less prompting. More finished work.

MXterm Pro: a native terminal for AI coding agents (Claude Code, Codex, Gemini CLI), SSH and SFTP, with a Control Room, git worktree per agent and a Token Meter. macOS, Windows, Linux.

Connect any LLM to a live OpenTTD game via MCP. Real geography from Google Maps. Reference Squirrel AI inside the game. MIT.

Repository diagrams that cite their sources. Every edge names the file, line and commit it came from — and every file it could not read is listed by name. Six languages, zero dependencies, one self-contained HTML file. An agent skill for Claude Code, Cursor and Gemini CLI.

Catch when your agent skills, instructions, and rules stop triggering — or stop working.

Lean MIT skill stack for coding agents: plans, TDD, verify-loop, constraints, review, PRs, and ticket pipelines.

AI-powered test generation assistant that converts JIRA requirements into structured test cases and executable Playwright + TypeScript E2E tests, with framework-aware code generation and automated quality validation.

An agent skill to diagnose and fix software bugs across languages and frameworks using root cause analysis, targeted fixes, and regression checks. Install with npx skills for Claude Code, Codex, GitHub Copilot CLI, OpenClaw, Hermes Agent, and more

ERC-8350 (Draft, merged into ethereum/ERCs): Agent Memory State Registry — authorized commitments to private agent memory state transitions. Reference implementation + golden vectors + Sepolia deployment.

Fast Windows file discovery for Codex, Claude Code, and compatible AI agents using Voidtools Everything CLI.

A curated list of 250+ tools for Automated AI/ML - from hyperparameter optimization to autonomous AI agents.

Evidence gates and session continuity for AI coding agents

Lets coding agents (Claude Code, Codex, pi and others) see and operate macOS apps: screenshots, the accessibility tree, clicks, typing, menus and windows.

Skills that teach your coding agent to turn an iOS Figma design into real SwiftUI — and keep the two in sync

Generate a plain-English receipt of mint authorities, supply, metadata, and risks.

Learn to inspect external sources before AI agents use them: inert exercises, static risk receipts, and clear boundaries before execution.

Package a tiny HTML game, rules, and provenance as a transaction-sized artifact.

Text/image/sketch to a textured, rig-ready 3D GLB in one call. The three.ws Forge generation SDK: free TRELLIS lane, pay-per-call USDC tiers over x402, and auto-rigging.

The three.ws agent engine: a plan/execute decision loop with human-approval gates, a seven-layer GuardChain (security blacklist, intervention, capability, permission, trade guard, spend envelope, x402 budget), a tamper-evident hash-chained action ledger, and a multi-step transaction pipeline with rollback.

One command to connect any MCP client to three.ws: sign in with OAuth or an API key, write the three.ws MCP servers into Claude Code, Claude Desktop, Cursor, Windsurf, VS Code, IBM Bob, Codex, Gemini CLI or Hermes, print the connector fields for Grok Bot, and verify each with a live tools/list.

DeepSeek Harness plugin for agent workspace hygiene: artifact value assessment, metadata indexing, and auditable cleanup.

🔄 Transform any USB keyboard into a Bluetooth device with ESP32-S3, allowing seamless connections to multiple gadgets. Enjoy universal compatibility and low latency.

One source of truth for every AI coding agent. The standards, decisions, specs and plans of your organization, written into the AGENTS.md of every repository. Claude Code, Codex, Cursor and GitHub Copilot already read it.

Verified HTTP, SOCKS4 and SOCKS5 proxy data for developers and AI agents. JSON, CSV, TXT and API access with check timestamps, country, anonymity, latency and uptime.

A 3D office that mirrors your Claude Code sessions; prompt, approve and follow them from any device over Tailscale.

A local SVG logo editor for fine-tuning designs by hand or with Codex and Claude Code.

An Agent Skill that lays out PDFs page by page: full-bleed reports, proposals and decks with vector text, six built-in themes and a quality pass the agent runs on its own work before delivering. Claude Code, Codex, Cursor, Gemini CLI.

CLEAVE: skills-based CTF agents for mechanics-first challenge triage and exploit orchestration.

herdr mobile & web client that shows what your Claude Code mods report — answer your agents from your phone, over your own Tailscale

Claude Code skills to build a SaaS or ecommerce in Argentina: spec-driven dev, test-first debugging, security audits, safe migrations, UI checks, plus ARCA, Rebill, Mercado Pago and Correo Argentino.

Native iOS client for DeepSeek Harness (dsh). Answer your agent from your phone — the relay only ever sees ciphertext.

Curated open-source tools and building blocks for GTM engineering: prospecting, enrichment, SEO, research agents and outreach.

Reverse-engineer closed-source desktop software into CLI tools for AI agents. Verified CapCut, Axure RP and Shapr3D workflows.

Автоматическое обезличивание персональных данных (ПД) по 152-ФЗ. Готовый MCP-сервер и API для безопасной работы ИИ-агентов, n8n и LLM.

给 DSH 模型每轮回填宿主实测的上下文占用，并阻止"上下文快用尽"的错判被压缩摘要继承。

Agents are the newest form of leverage. Skills + blueprints that give yours something real to work with: live social data, people + company enrichment, a workspace Postgres, schedules, shareable dashboards. Solo founder, big company, day job, or your personal life: one install, any agent.

Free, open source video editor for Reels, TikTok and Shorts. A CapCut alternative for Mac, Windows and Linux with local Whisper captions and AI agents over MCP.

An agentic scriptorium for scholarly writing — coordinated AI capabilities for manuscripts, grants, and reviews.

Local desktop panel that turns your AI coding sessions into an isometric dwarf mining colony - watch, message, and manage your agents in real time.

A super simple, agent-friendly daily planner. MCP for Claude, Cursor, Codex, ChatGPT, Grok. Linear + Frame Dashboard.

Tap the screen, it gets fixed. Anyone at your company can build features, modules and fixes from inside your app; an AI coding agent turns each request into a PR.

MCP server for AI agents to connect and interact with remote servers over SSH.

Native Mac app that turns your Claude Code and Codex history into something you can read, search and see. Everything stays on your Mac.

售前工程师的本地 AI 工作台:客户决策链、商机流水线、交付物生成、知识库 —— 数据是你磁盘上的 Markdown

Agent-first display server: a small trusted core speaking a capability-native protocol, with every legacy app confined to its own per-app shim. Humans and AI agents operate the same GUIs under revocable, capability-scoped authorization.

Codex skill for approval-led experiment repositories: connect reports, configurations, raw evidence, and review decisions for delegated research.

Local two-way messaging between Codex Desktop and Claude Desktop Code on Windows.

Pre-print simulator for Bambu Lab 3D printers: printability, stability, strength (FEA), physics and robot/mechanism simulation — web app + MCP server for AI agents

Client-side HTTPS egress proxy for TokenSaver — MITM capture for Claude Code / Cursor / agents.

Native macOS approval panel for every Claude Code, Codex, Cursor and Antigravity session.

Human factors research translated into design standards, code examples, and an MCP server for AI-assisted validation.

On-device Retrieval-Augmented Generation using Qdrant Edge, Qwen3 embeddings, and Gemma4.

The comprehensive directory of the AI agent ecosystem: frameworks, platforms, coding agents, computer-use agents, orchestration, protocols, evals, memory, observability, marketplaces, security, and sandboxes.

Know what your AI agent built, and where it just used its defaults. A skill + plugin for Claude Code, Codex and OpenCode that briefs you on decisions, provenance, and failure modes so you can pass the whiteboard defense.

Create a human-readable receipt for agent deployment, authorities, and endpoints.

Create a compact, independently verifiable bragging-rights timestamp payload.

Compare Ethereum blob, calldata, and Solana v1 placement for small cultural artifacts.

A city-builder where the cities are your git repositories and the villagers are your AI coding agents.

Automate code review workflows with advanced AI analysis. Get intelligent insights on pull requests, track coding activity, and improve code quality at scale.

DeepSeek Harness port of Claude Code Game Studios (CCGS v1.1.2, MIT) — one-package install: hook bridge (9 events), studio status line, self-installing framework payload (76 skills, 49 agents, 13 rules).

ClaudeRabbit is a free, open-source security product protecting the open-source community from malware. We clone any public GitHub repo into an isolated sandbox, run it, and return one honest safety score.

Curated landscape of 500 agent runtimes, sandboxes, browser agents, tool protocols, memory, observability, and compute platforms.

Kubernetes operator that manages AI agents as first-class resources

Linked memory records you own. Tamper-evident memory with a resume phrase as the key. Local-first, zero dependencies.

The structure a project grows on: a free starter template for AI-assisted development. One AGENTS.md every tool reads (Claude Code, Cursor, Grok Build, Codex, Gemini CLI, Copilot), TDD + bug-fix methodology, secret guardrails, optional local knowledge base. Built for founders new to vibe-coding.

Autonomous AI agent that searches PubMed, Europe PMC, OpenAlex and arXiv, ranks papers for any research field with embeddings and an LLM, and emails a curated digest

Fine-tune, calibrate and gate Laya (open-weights System One decision model) on your own decisions, then guard production agents with it

Headless Teneo Beacon for Raspberry Pi, Linux servers, NAS and macOS. One-line install, QR pairing, runs 24/7 as a service and earns Fragments.

我蒸馏了 645 个 B 站勇哥说餐饮真实视频与血泪案例，为你打造了这款餐饮创业防骗 AI Skill。专注破解加盟割韭菜、选址连环套、转让接盘侠三大陷阱，帮你在开店前守住本钱。

Encode compact generative art parameters into EVM calldata and compare Solana v1 bytes.

Visualize up to 64 inline addresses and detect duplicates rejected by v1 sanitization.

Model total v1 priority fees for bandwidth-heavy transactions without instruction scanning.

Package executable creative-code sketches with an immutable source hash.

Decode the v1 envelope discriminator, config mask, addresses, instructions, and signatures.

Engineering-discipline skills for AI coding agents — /think, /tdd, /review, /debug as composable skill documents. Skill-spec compliant, no runtime: the documents are the product.

The agent hub on Stellar: create AI agents with their own on-chain identity and wallet, find them, and pay them in USDC.

A++ clean code from your AI agents. Zero effort for you, fewer tokens for them. Gates for TypeScript, Python, Rust, UI, databases and tests.

EML (Engram Protocol): a typed memory store for AI agents, and the EML Cockpit — a local app that runs agent CLI sessions next to your own brain.

🧠 Persistent local-first memory for AI coding agents — MCP server + CLI for Claude Code & Cursor. SQLite FTS5, secret redaction, project isolation.

GitHub-native development workflows for Codex, with reusable Skills, durable Issue/PR records, and human oversight.

Agent Jobs: your main Claude Code session talks, decides and checks; everything longer than a few steps runs as a background claude -p job – own worktree, hard limits, guarded, one-line result. Cheaper context, free session.