Code agents

Accessibility Skills (accessibility-skills): AI skills that plan, review, test, and audit web accessibility (WCAG 2.2) — Accessibility Planner, Critic, Tester, Perspective Audit, and Bug Reporting for Claude Code, Codex, and local models, with cross-model benchmarks.

Open-source AI agent toolkit for Roblox Studio

Behavioral anomaly detection and runtime observability for AI agents

Local-first, policy-guarded AI operator layer with governed execution, mission continuity, observability, and receipts.

Runtime security & governance control plane for AI agents. Enforces per-tool authorization, OPA policy, behavioral risk scoring, blast-radius analysis, autonomy contracts, and tenant kill-switch. Every decision is ed25519-signed and Merkle-anchored for tamper-evident audit. 12 microservices, sub-30ms p95, OWASP LLM Top-10 coverage.

Interrupt-aware MCP sidecar for long-running coding agents

Flexcompute AI plugin marketplace for Claude Code, Codex, and GitHub Copilot CLI.

Agentic lead-generation workflow built on Xpoz MCP + Claude Code — finds buying-intent leads from Twitter/X conversations

Benchmark harness for AI-generated 3D browser games — same brief, same tools, 8 models, 24 playable builds.

Audit and prune the standing rules a project operates under (CLAUDE.md, memory, AGENTS.md, ADRs, conventions) in one whole-corpus pass.

Govern, audit, and ship AI-agent skills with policy checks, evidence, SBOM, and MCP.

Deterministic finish-line gates for AI coding agents. A model-independent evaluator that blocks commit/publish until your definition-of-done passes. Works in opencode (any model), Claude Code, pre-commit and CI.

Safe, quality-gated model+effort router for Claude Code - delegates cheap work down to save tokens, escalates hard/sensitive work up through a verify staircase, never-down gate, and fail-safe Lock, up to a consent-gated Fable 5 top rung. Not the cheapest router; the safest one inside Claude.

A curated collection of Agent Skills for Claude Code and compatible agents.

High-fidelity Word (.docx) to Markdown converter. Preserves tables (vMerge), footnotes, field codes, bibliography, lists, and images where Pandoc and others fall short.

本地优先的文档数据管理工具：把大规模语料整理成带稳定地址的条件树，精确定位片段且可回溯出处；关键词 + bge-m3 语义检索，经 MCP 开放给外部 AI agent。

Discover, Verify, Transact, and Rate an Unknown AI Agent. Agent discovery, trust verification, and capability routing powered by Aidress — the coordination registry for autonomous AI agents. Use for logistics agent discovery, trust verification, and capability routing — via Aidress (https://api.aidress.ai).

Free AI agent skills for business automation. Sales, content, ops, growth. Compatible with Claude Code, Codex, Gemini CLI, Cursor.

Open Agent Skill for Codex, Claude Code, and GitHub Copilot that turns rough ideas into researched decisions and practical outputs.

FireGuard hooks for AI agents: protect Cursor, Claude Code, Windsurf, Devin, and Copilot Studio from secret leaks, unsafe commands, and prompt-borne threats before they execute.

In-session compliance enforcement skill for Claude Code (paper Sec 6 dogfood). Per-user adaptive threshold + chaos-tested install/doctor/uninstall.

makes your claude code faster and more efficient, cutting token usage by up to 70% through orchestration, right-sizing, warm cache reuse, and more.

An agent for your whole life—built with contestable memory, explicit authority, multi-channel access, and user-owned data.

Curated Skills, MCP servers, and tools for AI agents.

GCF Rust implementation. 100% LLM comprehension on every frontier model. 50-92% fewer tokens than JSON. 43B+ round-trips verified. Zero dependencies.

Claude Code plugin: spec-driven agentic coding with cross-model review gates (Claude builds, Codex reviews), a self-hardening findings ledger, and repo-enforced quality.

Self-hosted remote workbench for your projects — agent terminals, files, git, kanban, mindmap and scheduling, reachable from anywhere.

Native Claude Code + Codex skills for AI security review. Also ships instruction files for Cursor, Copilot, and Windsurf. Checks 33 risks: OWASP LLM Top 10 2025 · applied RAG/MCP/agent checks · OWASP Agentic 2026.

Lighthouse for AI shopping: audit any product page for AI-agent readiness. 0-100 score, evidence-backed fixes, offline, no API key. (CatalogReady)

AI-Powered Documentation Generator for Claude Code

100+ agentic skills and automated workflows for Product Managers. From idea discovery to vibe coding.

Runtime-evidence debugging for coding agents: session-scoped probes, bounded logs/query CLI, and an agent skill.

Keep long-running repository work coherent and easy to resume.

Claude Code skill that clones Fable 5's reasoning habits — the playbook Fable wrote about itself. A/B validated: 20/20 convergence, blind-judge parity. One-line install.

Temporal debugging skill - time-travel through git history to debug bugs in their original context

Graph-powered code intelligence for AI agents — a Rust engine that maps your whole codebase into a knowledge graph and serves it over MCP, so Claude/Codex/Cursor answer structural questions with ~40x less context.

AI-powered triage for SAST findings.

Stable Android device automation for AI agents. Uses Accessibility Service + MCP protocol — no ADB, no OCR, just native OS APIs over WiFi.

A production-minded Codex agent harness for new and established repositories, with progressive skills, repeatable verification, safe upgrades, and security gates.

🩺 A meta-skill for Claude Code that gives your other skills a health checkup — diagnose-only, based on Matt Pocock's 'Building Great Agent Skills' method.

Trace — a local, background version tracker for multi-CLI-agent coding workflows (Claude Code, Codex, Cursor, ...). Content-addressed snapshots in SQLite, per-agent attribution, diff/restore/revert via an Electron console and an MCP server.

Fast local code intelligence for AI agents powered by AST Treesitter, SCIP, semantic search and a reranker

A highly configurable, self-evolving AI Agent operating system with plugin pipeline, multi-layer memory, hot‑reload, and MCP support. Built on FastAPI & React.高度可配置、自进化的智能体操作系统，支持插件化管道、多层记忆、热替换和 MCP 协议，基于 FastAPI 和 React 构建。

Governance, skills, and hooks stack for running Claude (or any capable LLM) as an autonomous senior engineer on any software project — frontend or backend, any language. MIT.

AI 🤖 automatically generates exploit PoCs for WordPress and WordPress plugins already patched! 🔥🔥🔥

Kit de qualidade de código para o ecossistema Delphi. Reúne modelos, automações de cobertura e diretrizes de Clean Code/TDD para alinhar desenvolvedores e agentes de IA na busca por conformidade estrita com o SonarQube (DelphiSonar)

Safety-first agentic coding framework. Three-layer safety pipeline (sandbox, YAML policies, human-in-the-loop approval) for AI coding agents. Pluggable runtimes (Claude Code, Codex), autonomous task orchestrator, full audit trail.

Agent skills built from Algorithms for Decision Making (Kochenderfer, Wheeler & Wray). Expected utility, value of information, bandits, robustness, belief filtering and game theory, as SKILL.md files with a zero-dependency calculator.

AI-powered interview preparation platform built with LangGraph, RAG, FastAPI, and MCP. Simulates adaptive technical interviews, evaluates answers, and generates personalized study insights.

The agent-native deploy platform you actually own — ship AI apps inside your own (or a client's) infrastructure, driven by your AI agent over MCP, with SSO + per-user audit. Self-hosted, open source; a deploy layer for forward-deployed engineers.

An LLM-maintained knowledge vault. Folders over agents: a Claude Code + Obsidian template for compounding what you learn about any fast-moving topic. Inspired by Jake Van Clief's ICM.

Claude Code plugin: quiets noisy Bash output (tests, builds, buildkite, docker, bazel, big git diffs) by redirecting to a temp log and surfacing only a summary.

Run investor relations like you have an IR team of ten. Open-source, local-first IR system for founders: cap table & SAFE math, financial dashboard, investor CRM, update composer, data-room playbooks — built to be operated with your AI agent.

Agent-native knowledge base of Apple WWDC sessions (2014-2026, Tech Talks, Meet with Apple): clean transcripts, structured metadata, code, and sosumi.ai/DocC doc links. Generated from Apple's public feeds.

Markdown protocol for multi-vendor AI agent teams — capability cards, bandit routing, and lessons that compound.

A flight deck for your AI agents — multi-agent terminal platform built on WezTerm + tmux + git worktrees. / 为 AI agent 时代的多 agent 驾驶舱，基于 WezTerm + tmux + git worktrees。

Agents are starting to run real business processes. flowproof tests them like anything else: record one real run, then replay it on every commit with zero LLM calls, asserting which tools were called, with which arguments, in which order, and which were not.

Open-source Codex workspace, skills and verifiable-document tooling for evidence-grounded legal work

Portable Agent Skill for small, verified coding-agent changes

Offline MCP server: an AI agent transpiles PineScript v6 → C++ and runs deterministic, TradingView-parity backtests locally — 309/311 strict on the open corpus, 396/396 on a closed 415-script community test, ~950k trades verified, 0 engine bugs. One Docker container, no API key; code & data stay on your machine.

Docker-MCP-Server - An MCP server covering the full management surface of Docker. Manage, maintain and audit multiple docker environments with ease.

Drive your real, logged-in Chrome from an AI agent (Claude Code / Codex) over MCP - 55 tools for browsing, DevTools-grade network capture, and web-security testing.

Policy you install, not a framework you run: an orchestrator–worker discipline for Claude Code. One Node installer, identical on macOS, Windows, Linux, and servers. Built for codebases that get audited.

REST API for MetaTrader 5 with agent skills

MCP server for GitLab API — 76 tools for AI assistants (Claude, Cursor, VS Code). Projects, MRs, pipelines, CI/CD, approvals, issues, code review.

Local-first kanban board where Claude Code agents work your cards — dependency chains with merge gates, an AI manager that reviews real diffs and merges green PRs, group lanes, CI-gated. One runtime dependency.

The open interoperability layer for AI agents. Connect your favorite agents and harnesses. Work together.

Odyssey-AI is an intelligent travel assistant that provides personalized travel plans, real-world photos, and AI-generated visuals for an immersive planning experience.

Understand core techniques of harness-engineering by building a mini Claude Code

MCP server for Okta with agentic API discovery and guided code execution

PromptKit workflows for coding agents, the command line, and Go — agent skills plus deterministic, offline prompt discovery and assembly.

MCP server for the Rakuten Web Service API. 28 read-only tools across Ichiba (e-commerce), Books, Travel, Recipe, Kobo, and GORA (golf). Bilingual EN/JA tool descriptions.

Node.js TypeScript CLI for discovering, staging, activating, and wiring reusable AI-agent assets across VS Code/Copilot, OpenCode, Cursor, Zed, Claude Code, and Pi.

Self-hosted lifecycle platform for software powered by Claude Code. Configurable pipelines, devices you control, no credentials on the server. Apache-2.0.

Self-hosted coordination and observability hub for AI coding agents

Publish AI-generated content in seconds

Run a Karpathy-style planner/generator/evaluator agent loop in your repo: a manager negotiates a testable contract, then drives a resumable background runner.

Keep one work frame across Claude, Codex, and Gemini.

AI Project Starter: documentation-first starter kit for AI-coded projects, project context, guardrails, and agent instruction files.

A Claude Code plugin that offloads token-heavy, self-contained tasks to local CLI backends.

Your agents change the code, then explain why. This skill makes them bet on the result first.

One fuel gauge for every AI coding harness you actually use.

AI agent pipeline: drop requirements into Feishu, agents code/test/deliver in isolated git worktrees. Single-binary Go daemon

Forge rough ideas into drift-proof execution contracts for AI coding agents. 6-step pipeline, 16 agents, 8 skills, lifecycle hooks, 5 tech presets.

Zero-error LLM execution via SPRT voting. Rust library and MCP server implementing the MAKER algorithm for mathematically-grounded error correction in long-horizon AI agent tasks. Research experiment based on arXiv:2511.09030

A framework-agnostic voice agent library for voice-controlled UI interactions, powered by OpenAI's Realtime API.

Loop engineering for Power BI: reload a PBIP so Desktop re-reads TMDL/PBIR changed on disk, and let an AI see what rendered.

AI-Factory — autonomous pipeline that designs, builds, tests, and publishes products.

Enable AI agents to access and analyze real-time social media data from Twitter, Instagram, Reddit, and TikTok using standardized skills.

Lightweight dual-agent CLI + skills toolkit for Claude Code and Codex workflows.

A fast, OS-sandboxed Model Context Protocol gateway with an embedded Code Mode engine and shell-output token compression, in a single Rust binary.

🏖️🛡️ Keep your LLM's 💩 code where it belongs — in a sandbox.

Detailed guide covering OpenAI Responses API, built-in tools (web search, file search, computer use), Agents SDK, and observability tracing

Universal, cross-tool persistent memory bank for AI assistants (Claude Code, Cursor, Windsurf, OpenCode, Codex) powered by FastMCP, Gemini embeddings & Supabase pgvector.

Agentic security control plane for MCP and AI agent tool calls. MCP-native policy gateway with topology discovery and audit.

3D model generators with CadQuery for humans and LLM agents

Turn MEMORY.md into project-scoped, versioned memory for Codex, Claude Code, and MCP agents across sessions.

A dating app for coders where you pick your waifu—and she becomes your AI agent, living on your device, completing tasks, and interacting with you in her unique personality.

Harness-agnostic, deterministic, resumable multi-agent workflows — plain-JavaScript orchestration for fleets of coding agents, driveable from any MCP client (Codex, Claude Code, Cursor), with key-based resume, worktree isolation, personas, and a live viewer

Privacy-first iOS SSH terminal — agent- & multiplexer-agnostic (tmux, zellij, screen; Claude Code, Codex, aider, Hermes…). On-device dictation, Live Activities, zero data collection.