Code agents

How I scored top 10 on Optimization Arena negotiation + prediction market challenges by reverse-engineering the leaderboard via the public API

MCP server for the Linux desktop 🐧🤖

Watercooler: your team's shared reasoning layer for agentic coding. Preserve the "why" around your code.

Native macOS control panel for Flutter — for editors (Zed/Neovim/terminal) and agentic coding without a Flutter plugin

ChatGPT Codex — cloud-based AI coding agent with sandboxed execution, multi-file editing and automated test generation for software engineering tasks.

Your AI assistant on autopilot - A powerful automation framework for ChatGPT that simulates human-like interaction patterns. No API required!

AgentDesk – AI-Powered SaaS Platform for Business Chatbots and Automation

Official Roadmap for Catalyst – Track upcoming features, improvements, and releases. Submit feature requests and help shape the future of Catalyst

Autonomous delivery orchestrator for Claude Code — hand it a goal and a numbered Definition of Done, it runs the whole job to completion

Claude Code, freed from your terminal: parallel sessions on any device, scheduled jobs, PR reviews, and any model provider.

The WorkBoard for Claude Code and humans. Watch your work come to life — never lose an idea, never lose a workflow.

One-click unlock for Claude Code ultracode effort mode (xhigh + dynamic workflow orchestration). Version-adaptive patcher with daily auto-update cron. Supports 2.1.113+ including 2.1.156, 2.1.158 and newer.

Portable Ethereum identity for your AI agent. Its soul, memory, and skills live onchain via ERC-8004 + IPFS and snap back into any session.

No-code AI voice agent for n8n: calls your leads, qualifies them, pulls their email out of the conversation, and logs every transcript to Google Sheets with Slack alerts. Vapi, Twilio, OpenAI.

Structural security, governance and execution for AI agents: validate actions before they touch files, credentials, network, APIs, or money.

AI usage limits in your MacBook notch or menu bar — Claude, Codex, Gemini, OpenAI & more. Native Swift, local-first, zero telemetry.

Local-first, eval-first memory for long-horizon AI agents — no LLM at ingest. Python SDK + MCP server with source-traceable recall, belief revision, selective forgetting, and reproducible benchmarks.

Fast headless browser and native app automation CLI for AI coding agents — persistent Chromium via Playwright, camoufox stealth runtime, session multiplexing, MCP server, and Android/iOS/macOS app automation.

MCP server giving Claude live access to SAP systems via ADT — read, edit, search, test, and diff ABAP across landscapes without installing anything on the SAP stack.

MindWiki — 10 顶级专家组团的决策追问系统 Skills（张小龙、乔布斯、马斯克、张一鸣、芒格、王兴、雷军、梁宁、Naval、毛泽东）。带工具调用和自我验证测试。

Claude Code plugin that makes coding agents measurably cheaper over time: collect token costs, distill candidate rules, benchmark them on a frozen golden suite, and keep only rules that earn their context rent.

Evolving agent harnesses: a research program on how far N orchestrated calls of a small model can rival a frontier model. We evolve the harness (structure + prompts) with reflective optimizers + a verified-acceptance gate.

Open re-implementation of Claude Code's built-in Dynamic Workflow: a script-driven runner for agent-authored JavaScript workflows — agent(), parallel(), pipeline(), token budgets, git-worktree isolation, deterministic resume — plus pandacode, a pluggable executor for codex, claude, and domestic LLMs. Auto offline HTML execution-graph reports.

Prismatic — a highly customizable, zero-dependency status line for Claude Code. Truecolor gradient model pills, a color-coded context bar, and a live visual config editor.

Prompt optimizer for Claude Fable 5, built as a claude.ai Project — 43 source-backed rules, complexity routing, and API parameter recommendations, grounded exclusively in official Anthropic documentation.

Claude Code plugin marketplace — agentic-engineering (spec-driven shape→decide→execute→measure→eval with adversarial review) + github-keeper (audit/elevate READMEs and make a repo open-source-ready).

Zero-hallucination reference verification for Claude Code — live CrossRef/S2/PubMed verification, verbatim abstract traceability, retraction check

A self-correcting system of nested engineering loops: plan & execute repo work; safely implement audit findings; design greenfield & improve brownfield; learn from failures; re-verify until closure.

Eliminates generic AI slop and enforces premium UI/UX design via 10 sequential gates, ratio-based layout/composition audits, and open-design craft integration.

An unofficial orchestrator skill for Matt Pocock's agent skills. Routes you to the right skill based on your situation.

Generate schema-valid Paperclip agent-company bundles from a Markdown brief.

SkillBill packages your engineering judgment into governed skills, then uses a durable Kotlin runtime to ensure agents actually execute that process consistently, resumably, and with bounded context.

Open-source MCP server connecting Claude AI to Oracle EPM Cloud via REST APIs. Run business rules, export data, and automate month-end close through natural language. Full tutorials and articles at fmepm.com. Built by Fred Mamadjanov, Oracle ACE Pro.

Universal multi-agent infrastructure for AI coding assistants. 26 specialist agents, 19 enforced workflow skills, and a lead orchestrator — for Claude Code, Codex CLI, Cursor, and OpenCode.

[ICML 2026] Rule2DRC: Benchmarking LLM Agents for DRC Script Synthesis with Execution-Guided Test Generation

The tool for interactively reviewing an HTML plan for Claude Code

AI-assisted reverse engineering orchestration platform for IDA Pro and x64dbg with live synchronization and MCP agent workflows.

Open-source Human Design AI: BodyGraph charts, whole-chart talent and mission readings, DeepSeek consultation, FastAPI/React app, and Agent Skills. 开源人类图全盘解读。

LLM Council for Claude Code with zero API keys — parallel Claude subagents debate, anonymously peer-review, and synthesize an answer, all on your existing subscription

Run a fleet of AI coding agents (Claude Code, Codex, Aider) across many repos — Rust TUI, tmux-backed.

AI reasoning skills distilled from 4,665 real Claude Fable 5 chain-of-thought traces. Mathematically tuned. Grade A (100%) emulation accuracy against MiniMax M3.

Claude Code / Codex / Cursor AI 编程规则：Karpathy 编码准则中文版，减少错误假设、过度工程化和无关改动

A professional plain English style guide and agent skill that translates legalese and purges AI slop.

Universal AI Agent skill for downloading videos from YouTube/Bilibili/Douyin/Xiaohongshu/Kuaishou/Weibo/Twitter/Vimeo/TikTok. Works with Codex, Claude Code, Cursor, Cline.

Personal collection of Agent Skills for AI coding agents (Claude Code, Gemini CLI, Cursor, Codex, etc.)

An autonomous agentic pipeline that finds, proves, and patches real C memory-safety vulnerabilities end-to-end using a single 7B open-source LLM (Qwen2.5-Coder) on vLLM, with AddressSanitizer as a ground-truth oracle.

Detent is board-driven agentic work orchestration, shipped as a single Go binary — it watches a GitHub Projects board and runs each ready issue through an isolated Git worktree, a coding agent, your validation gate, and a serialized merge train, live on web and terminal dashboards.

An Always-On Autonomous AI Software Engineering Team built with LangGraph, FastAPI, and Next.js that operates natively within your GitHub repositories.

The first free, open-source, self-hosted MCP server for Google Merchant Center (GMC). 126 tools over Merchant API v1 — products, inventory, reports, promotions, returns, account config, audit log + rollback, dry-run. Free alternative to paid SaaS (Adzviser, Catchr, Windsor.ai, Pipedream).

A comprehensive Android staff management application featuring role-based access control (RBAC), real-time chat, and a RAG-based AI assistant powered by Gemini API.

YAML hook automation for the PI coding agent: tool guards, session hooks, prompts, notifications, and bash actions.

Your coding agent starts every session with amnesia — memo fixes that, 100% on your machine. Persistent memory for Claude Code, Codex, Cursor & any MCP client: Markdown source of truth, hybrid search (MLX/CPU + sqlite-vec), time-machine, contradiction radar, nightly self-optimization. No cloud, no keys.

AI agent with zero dependencies, bring your own bun

Engineering best practices that AI coding agents actually apply — 40+ skills, from secure coding and cloud to compliance and UX writing. BUILD + AUDIT modes. Measured: best-practice coverage 59% → 98% vs. the same model unguided.

Open-source AI interview copilot & meeting assistant. Real-time transcription, AI-suggested answers, screenshot reasoning. BYO API key, or run fully offline with Ollama. macOS + Windows.

gradio region:us

transformers gguf code granite text-generation dataset:bigcode/commitpackft

transformers safetensors llama text-generation code granite

gguf code quantized GGUF imatrix quantization

transformers safetensors qwen2 text-generation unsloth code

transformers safetensors qwen2 text-generation code codeqwen

gguf gemma4 coding code reasoning thinking

transformers gguf llama-cpp gguf-my-repo text-generation base_model:Qwen/Qwen3-Coder-30B-A3B-Instruct

Native macOS capture-to-outcome workspace: screenshots, voice, AI-reviewed planning, local semantic memory, and optional one-way Apple/Google exits.

A warn-only runtime layer for Claude Code agents. Keeps feature state, scope and handoffs on disk so they survive compaction, and requires an evidence path before a done claim. Never blocks a tool call. Companion to superpowers; bash + git, no MCP.

transformers gguf qwen-coder unsloth MOE pruning

A self-hosted Slack/Telegram agent that extends itself by writing and deploying its own code at runtime.

A local transparent proxy for Claude Code that captures every Anthropic API call — extract the full system prompt, track token & cache usage, and analyze tool-call patterns from a CLI or live React dashboard.

67 hand-picked Thai-language Claude Skills for working professionals and students 🇹🇭 · Free, open source

See a game. Ship a game 🐾 An Agent Skill that clones casual web games from a single prompt — provably-correct levels, browser-verified, original-branded. Works with Claude Code, Codex & more.

A production-grade, microservice-driven exam management platform that uses LangGraph AI agents to generate exams from classroom materials and strict-RAG evaluation to grade subjective answers — all within a real-time, glassmorphic dashboard.

Agents answer. ZugaMind notices. A persistent cognition sidecar that wakes Claude Code, OpenClaw, Codex, or Hermes when something deserves attention — engineered Global Workspace Theory, zero dependencies.

Developer integration tools for adding TDM payments to AI agents, apps, and workflows. CLI, MCP server, and copy-paste examples

🌱 Open-source visual platform to build, run & ship AI agents and workflows — built-in RAG, real-time collaboration, self-hosted. No code required.

An AI client for MSX2 machines with a RAM Mapper and a UNAPI TCPIP device.

A workflow specification for autonomous agents

Run your own unrestricted 70B+ AI locally · No limits · No filters · One-time payment

censuschat aims to let you ask questions about U.S. Census data in natural language. More specifically, a large language model takes queries about Census data and translates them into tidycensus R code. It uses the ellmer R package under the hood.

Windows에서 더블클릭 한 번으로 Claude Code를 설치·실행·업데이트·제거까지 관리하는 한국어 원클릭 키트 (비공식, PowerShell 메뉴)

Personal Claude Code configuration versioned inside ~/.claude — agents, skills, hooks, settings, and statusline (reference, not a starter template)

Mission control for your Claude Code sessions — a 100% local dashboard to see, resume, and drive every session on your Mac.

Система автоматизированной проверки устных ответов с Yandex SpeechKit, YandexGPT и поиском по учебным материалам.

探客AI ai-tungke 外贸智能获客平台：基于 Google 地图 + 高德地图的深度数据挖掘系统，遍历式地点检索 × AI 联网搜索 × 行业数据综合评估，让外贸 B2B 获客 lead generation 更精准高效

Loop-core multi-agent coding tool that orchestrates work through plan→implement→review cycles. The orchestrator is read-only — it searches project memory, enforces custom rules, and delegates all code changes to parallel swarm sub-agents. Built on Kimi Code CLI (MIT).

No API keys required. Transform the native ChatGPT, Claude, and Gemini web interfaces into fully autonomous local coding agents.

A local workbench for building, running and debugging AI agents.

Decrypt WeChat databases on macOS by extracting encryption keys to access and export chat records with support for searching and AI query integration.

The whoburnedmore CLI — count what your AI coding agents really burned and submit to the public token leaderboard. Source of the npm package `whoburnedmore`.

See where your Claude Code tokens go — cost per call scales 4.3x with context size. Hooks + SQLite + dashboard + MCP server for context window forensics.

OpenCode — an open-source, community-built VS Code client for the opencode CLI agent. Multi-tab AI chat, side-by-side diffs, checkpoints, cost tracking, and 75+ providers. Free (pay only your AI provider).

Repairs predictable tool-call mistakes that often derail open weight models. Supports integration with other extensions' custom tools.

Production-shaped boilerplate for AI agent web apps: Rust/Axum API, Python LangChain + Celery workers, Vue 3 SPA. Hexagonal architecture, TDD, JWT + refresh rotation, reliable job lifecycle with retries, rate limiting, structured logs, Docker & CI/CD

The brain is a git repo. Memories are markdown files with YAML front-matter, committed like any other file. Git history is the audit trail; pull requests are the approval gate — nothing is written to the shared brain without human review.

Open-source, self-hosted platform to build, host, and chat with AI agents — LLM chat bots and image/video generation agents — from a single interface. Bring your own API keys.

Portable AI Agent Workbench for safer DNF/DFO Script.pvf inspection and editing. 自包含的 DNF PVF AI 工作台，支持只读分析与受控修改。

Production-grade software delivery for coding-agent teams on Multica — dynamic planning, deterministic execution, verifiable delivery.

DVLAA (Damn Vulnerable LLM and Agent Application) is a local LLM and Agent security range for OWASP LLM Top 10 and Agent security training.

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

咨询风格 PPT 生成 Skill：core 共享逻辑 + Claude/Codex 双平台适配，默认全原生可编辑，HTML 样张零图像依赖，12 视觉风格 + 10 场景框架

Development discipline + multi-model subagents for OpenAI Codex — 13 dev skills, PABCD workflow, repo map, recall, skill search. Single plugin, no external harness.