Code agents

Local-first CLI that audits Claude Code transcripts for spend by model, avoidable context-carry waste, always-on config cost, and AI-fluency signals. Runs fully offline by default; opt-in tiers gate any egress.

Claude Code skill: pick the right loop type (/goal, /loop, /schedule) and write stop conditions that actually hold. From the Claude Code team's loops taxonomy + real overnight autonomous runs.

macOS menu-bar app that watches your Claude & Codex usage and auto-switches accounts before you hit a rate limit.

A cute desktop pet with a live Claude token usage meter - session & weekly rate limits in a thought cloud above your Windows taskbar. Free, open source, one-press install.

Self-hosted AI gateway for agents — route providers, virtual models, limits, and usage from one console.

Lunar Vault is a self-contained memory infrastructure for Codex. Features zero-dependency ONNX embeddings, advanced reranking, and secure one-click installation without leaking personal data

Kotlin Multiplatform semantic cache for LLM calls: eleven guards against false cache hits, measured on blind corpora, coroutine-first.

Corpus poisoning attacks against RAG-based cybersecurity/IoT threat-intelligence assistants, with a lightweight, edge-deployable provenance defense. Evaluated on TF-IDF and dense-embedding retrievers over a CVE/MITRE ATT&CK/vendor-advisory knowledge base.

Turn any website into a portable, agent-ready Open Knowledge Format (OKF) bundle — no LLM required. Crawl, sync, chat offline, and serve docs to coding agents over MCP.

LoRA fine-tuning of bi-encoder retrievers with hard negatives and cross-encoder distillation, evaluated on NFCorpus.

Local-first shared memory for AI agents: persistent AI memory across Claude Code, Codex, Hermes Desktop, and any MCP client. Automatic record, recall, consolidation, forgetting, and backup. Self-hosted on Qdrant + LlamaIndex; one-command install.

Detect prompt regressions before they reach production — per-category accuracy scoring, deterministic validation, and False Improvement detection. Pure Python, zero dependencies.

Full-stack AI English coaching app built with React, Node.js/TypeScript & Express 5. Features Gemini-powered level-aware chat, WebSocket voice coaching, PostgreSQL + pgvector RAG help bot, and vocabulary bank.

A modular, local AI companion featuring a RAG pipeline with FAISS and SentenceTransformers for semantic long-term memory. Powered by GGUF models via llama-cpp-python

⚡ Blazing-fast GPU vector search for PostgreSQL powered by NVIDIA cuVS.

Code Navigation for AI

Molavi Agent Skills: Curated AI agent skills and MCP configs for Antigravity, Cursor, Codex & Claude — by Taghi Molavi

mlx safetensors kimi_k3 moe reap pruned

gguf code agent agentic-coding moe coding

原生 macOS AI 安全智能体：漏洞情报、知识图谱、多语言代码审计与报告交付 | SwiftUI + FastAPI + LangGraph

Turn AI agents into team members: assign objectives with a definition of done, talk in channels, and review every LLM call — self-hosted. Claude Code and OpenAI Codex runners included.

AI Business OS: multi-agent platform with RAG, OCR, speech & workflow automation. FastAPI + LangGraph + Celery + Flutter. Production-grade, Dockerized

Multi-Agent Coding Harness with pluggable AI-SDLC: switch sessions across claude, codex, gemini, export/share coding sessions, save tokens, multi-frontend (TUI · web · Telegram).

An autonomous AI chatbot built for WhatsApp using Node.js, Baileys, and the Vercel AI SDK powered by OpenAI. Designed to act like an active, context-aware participant in group chats and direct messages, Quart goes far beyond traditional command-and-response bots by actively listening and engaging in human conversations.

Paste clipboard images into Claude Code on Windows - fixes the broken Ctrl+V image paste. Converts any screenshot or clipboard image into a file path Claude Code auto-attaches.

Connect your Facebook Ads, Pages, Instagram and lead forms to Claude Code. Ask about spend, cost per lead and your best-performing ad in plain English. Meta Marketing API via curl, no MCP server required. Built by Irfad Imtiaz at My Legal Academy.

Open-source stock breakout signals: 8-source harvest, AI scoring, pump-and-dump filter, LightGBM backtesting on historical scans. Next.js, PostgreSQL, Docker self-host, agent API & x402.

A curated list of AI-native products: apps and tools where the LLM is the product itself, not a bolted-on feature.

Cẩm nang thực chiến Mastering Claude Code (tiếng Việt): 16 chương biên soạn từ tài liệu chính thức của Anthropic — CLI, workflows, MCP, subagents, skills, hooks, plugins, CI/CD & tối ưu context.

See where your Claude Code tokens actually go. Profiles the session logs Claude Code already writes, scores each project's context hygiene 0–100, and generates fixes backed by your own measured numbers. Local, zero deps, no telemetry.

Ragworks — a self-hosted, low-code workbench for building transparent, traceable RAG pipelines. Wire up ingestion and retrieval visually, then see exactly what every run did.

Long-term memory for AI agents on plain Supabase Postgres. Hybrid vector + full-text + entity recall with RRF, temporal validity, write-time dedup. You own the SQL and the rows.

Multi-agent skill for Claude Code that delegates tasks to Google Antigravity CLI — web search, image gen, codebase analysis, 40+ science databases

Official command-line and agent integration toolkit for Chat2DB Community, Pro, and Local: app lifecycle, datasources, SQL, and MCP.

A visual, local-first desktop shell for Oh My Pi — sessions, tools, models, permissions, and long-running agent work.

AI Agent 免 Root 免电脑永久控制 Android — 无线调试做跳板，adb tcpip 5555伪装USB模式，127.0.0.1回环自连

Multi-account Gmail MCP server on Cloudflare Workers — one deployment, one Google sign-in per connection, any MCP client

Turn your coding agent (Claude Code / Codex / Cursor) into a personal English tutor with persistent memory & spaced repetition — no quizzes, learn through conversation. 把 AI 编程助手变成有长期记忆的私人英语教练

MCP server that delegates Claude Code subagents to local models (LiteLLM/llama.cpp/Ollama), DeepSeek, AWS Bedrock, or any OpenAI/Anthropic-compatible backend — without losing the Claude Code orchestrator session.

An AI SRE for Kubernetes: connects to your cluster, investigates with real tools, explains the root cause in plain English, and executes the fix only after you approve it. Peer-reviewed — Journal of Grid Computing, 2026.

The open-source AI firewall for enterprise MongoDB databases. Safely execute live AI agent queries with zero sensitive data payload egress to LLMs.

Autonomous AI security engineer — multi-agent swarm that scans, exploits, fixes, verifies, and attests. Open source (Apache 2.0).

Multi-agent debate framework for long-form complex research. 11 AI departments cross-validate every claim with confidence annotations. Not one LLM — an adversarial committee. 共识管线：多智能体学术辩论框架，11个AI部门交叉验证，长线复杂任务一站解决。

Run Claude 3.5 Sonnet and Opus models locally with this desktop coding assistant for Windows, macOS, and Linux.

29 free Claude Skills for SEO and AI search. Built by Hawk Academy + StudioHawk. Drop into Claude Desktop and run.

Claude Code skill for building personal brands that convert customers, not just rack up views. Stage-aware playbooks, 21 worksheets, 40+ frameworks.

Remove Claude/AI attribution from a GitHub repo in one command clean commit history, force-push, and refresh the Contributors graph so @claude disappears.

A Rust SDK for Anthropic's Claude API that won't panic on you — no unwrap/expect/panic, denied at compile time. Typed structured output and tools derived from Rust structs, streaming, thinking, caching, batches.

Native Rust, single-binary MLX inference server for Apple Silicon. OpenAI/Anthropic-compatible LLM serving — text, vision, audio, embeddings — with the widest weight×KV quantization matrix of any MLX server. No Python, no GGUF.

Kotlin Multiplatform engine for running Gemma LLMs on-device on Android via LiteRT-LM — stateful KV-cache chat sessions, resumable model management, function calling. Includes NativeLM, a private Local AI chat app. AGPL-3.0 / commercial.

WardenIQ is an open-source AI test intelligence platform that turns requirements, docs, and code changes into test cases, test plans, coverage insights, and release test cycles for smarter QA.

中文 AI Skills 库：163 个技能包（63 中文 / 12 个本站自写），覆盖 PPT 18、Agent 调研 3、3D 6、游戏 5 | 每天自动复检，运行记录公开

Production-grade Agentic Harness for autonomous AI agents. Features MCP server integration, cost-controlled DAG workflows, and actor-critic loops. Optimized for Claude Code, GPT-4o, and local LLMs. The infrastructure for reliable Agentic AI.

A governed multi-agent software development Skill for Codex

Claude Agent Skill that eliminates generic AI-generated design. Stop your AI shipping the same purple-gradient, Inter-font, three-feature-card website as everyone else.

132 Claude Code tips from Boris Cherny (Head of Claude Code, Anthropic) — curated, German-language, auto-updated 2-3x daily. One-shot Implementation Guide for Claude Code auto-setup. Visual dashboard with filters & timeline.

An autonomous LLM paper-trading agent that dreams, remembers, and reviews its own trades

gguf llama.cpp gemma python code reasoning

Your Claude Code session sees the 5-hour limit coming, checkpoints its work, and resumes itself when the window resets. Hook-based, stdlib-only, macOS/Linux/Windows.

AetnaMem /ET-nuh-mem/ is agent Memory delivers fully local long-term auditable, proof-ready memory for AI Agents via a 4-tier progressive pipeline, with zero external API dependencies, it supports OpenClaw, install it via pip.

Job hunting beyond Thunderdome — a local-first, open-source toolkit that finds roles without the LeetCode/AI-screener gauntlet. Runs your whole search from the CLI, driven by an AI agent.

A local AI OS that keeps the apps you describe running, and acts when something happens. Your machine, your data, your accounts. MIT licensed.

A visual manager for fleets of Claude Code and Codex agents — real terminals, live context and cost, agents that message each other, across your Mac and any Linux box.

把 Mirasim 内置的 Anthropic 反代接给独立 claude CLI 和 Zed ACP agent — 动态端口发现，不复制凭证

Desktop client for the Asale token-sharing network: sell idle Claude, ChatGPT/Codex, Gemini, Kimi and Grok subscription quota for USDT, or buy capacity the moment you hit a rate limit.

Curso de AI Engineering en Rust para Jeresoft Academy: RAG, herramientas, agentes, seguridad y evaluación con evidencia reproducible.

A deep, multi-angle, adversarially-refuted code review skill for Claude Code, Cursor, and other coding agents

claude2codex — bridge AI coding sessions between OpenAI Codex CLI and Anthropic Claude Code CLI. Convert JSONL sessions bidirectionally and resume where you left off.

Scales system-prompt guidance to the model reading it: an oh-my-pi plugin that appends a capability-tuned directive block chosen by model family, version, and tier.

A brain for your mail: MailFathom turns IMAP mailboxes into a self-hosted, AI-native service. Mail synchronizes into your own PostgreSQL, is indexed for search and retrieval, and is served to AI agents over the Model Context Protocol. Semantic retrieval, answering, and gated write tools. .NET 10, Apache-2.0.

Browse, search & replay every Claude Code session on your Mac. Local-first session viewer, timeline scrubber & cost tracker (ccusage alternative).

mlx safetensors qwen3_5 qwen3_6 token-efficient efficient-thinking

Run AI agents in the Polkadot app. E2E-encrypted and private.

Multi-agent LLM trading system implementing arXiv:2510.04787: agents develop grid-trading strategies offline and a low-latency bot executes them on Binance spot. Research implementation under active correction, not for live capital.

Give your agent a memory it can prove. One Python file: local SQLite vault, hybrid FTS5 + vector retrieval, Cloudflare-aware fetching, and a research loop whose [V]/[E]/[H] citations are machine-verified by CLI, not prompt-enforced.

Fast, idiomatic Go MCP server for web search, image search, and page scraping to Markdown. Runs locally, no API keys, free.

A native Rust multi-agent operating system and swarm workspace — GPUI desktop chamber, headless gluon daemon, git-worktree isolation and an automatic merge gate.

🦙 chat-o-llama: A lightweight, modern web interface for AI conversations with support for both Ollama and llama.cpp backends. Features persistent conversation management, real-time backend switching, intelligent context compression, and a clean responsive UI.

A Telegram bot that drives four local coding CLIs — Claude Code, Codex, Pi, and Gemini — from your phone. macOS daemon with per-chat per-agent sessions, a confirm gate for dangerous commands, response streaming, image attachments, auto-installed launchd, and corporate-TLS (Zscaler) auto-detection.

Isolated Claude Code instances to keep things separated and clean.

Open-source AI-native engineering plugins for Claude Code — project context, code review, security audit, task clarification, and more

🥈 2nd Prize Winner (Expert Nomination) — Multi-agent clinical AI platform for InterSystems IRIS with FHIR R4, Vector Search RAG, LangChain, live FHIR agents, and a no-code Agent Builder.

A multi-agent engineering crew for Claude Code — one manager plans and verifies, specialist agents read, write, test, research, and review. Hard role separation, cost-tiered models, custom gateway ready.

Deterministic control plane for Codex Agents with typed delegation, durable lifecycle receipts, and journaled writer isolation.

Git-backed message bus for AI coding agents — Claude Code, Cursor and Codex coordinate asynchronously, with no server

safetensors llama Arabic English LLM Decoder

gguf code qwen qwen-coder codeqwen text-generation

AI-generated animated explanations of code — paste any function, get a cinematic animation of what it does

Self-hosted personal AI knowledge brain -- hybrid (vector + keyword + entity-graph) search over your notes and code, served to any MCP agent (Claude Code, Cursor, Codex). One AWS account, zero telemetry, MIT.

把学校规章、培养方案和公告变成能提问、可追溯到原始文件与页码的大学 RAG 知识库。

UniSearch 是一款基于 AI 的多平台公开内容采集与调研桌面工具。 通过自然语言描述调研目标，由 AI 自动规划任务流程，并发调度底层连接器进行公开数据采集与清洗。

Local-first semantic memory engine for AI agents - scores, trusts, and forgets so your agent stops making the same mistake twice.

FriendlyClaw 🦀 — Pro-Grade Autonomous Hive Operative. Multi-agent delegation, semantic RAG memory, and proactive system control.

Secure credential management for AI agents

The dashcam and emergency brake for AI agents. A security proxy that physically blocks rogue LLM commands and generates cryptographically proven audit trails for enterprise compliance.

🌟 Mindstride is a RAG-based chat assistant designed to support mental health, personal growth, and self-improvement. This project leverages advanced AI technologies to provide a personalized and insightful experience, drawing from the wisdom of over 70 books on mental well-being, personal development, and self-discovery.

🤖 一个基于本地 Ollama 模型的代码助手 CLI 工具。集成 LangChain.js，使用模型原生 tool calling 能力实现可靠的工具调用。

📝 Streamlit App : Weekly News Letter Crew AI Agents 🖋️

A centralized API management platform with user authentication, service versioning, collaborative API development and most valuable, automatic frontend code generation. Built with Python/Robyn backend and React/TypeScript frontend.

Multi-agent control plane for coding CLIs (Codex, OpenCode, CommandCode, Hermes) with one tmux session, one .logs/ source of truth, terminal + web frontends.