Havuz

DeepSeek Harness 基础工具增强：持久 bash、结构化 read、多模式 edit、原子 write、双引擎 grep/glob、图像直读，一个插件全覆盖

Serial port MCP server for Claude Code, Codex, Opencode and AI agents. Reliable UART/USB-serial access for embedded development.

aside-codemode, Aside içindeki JavaScript REPL üzerinden dosya araması ve browser işlerini tek çağrıda çalıştırıp modele yalnızca seçilmiş sonuçları döndürür.

Neden yükseliyor · 14 günlük yaşına karşılık 124 yıldız ve son 28 günde 287 commit görünmesi, hızlı bir başlangıç ve yoğun geliştirme faaliyetine işaret ediyor. README'deki ölçülebilir arama ve browser örnekleri de ilgiyi artırabilir; ancak yükselişin kesin nedeni bilinmiyor. Commit'lerin yalnızca 2 farklı yazardan gelmesi, faaliyetin geniş katılımla desteklendiğini göstermiyor.

Experimental scenario analysis for real-life events forecasting with Codex or Claude

Headless verification browser for AI coding agents: 35 ms one-call checks, pixel diffs, live human hand-off

From any API surface to verified AI agent tools: Go CLI + MCP + skills + SQLite + policy + proofs.

Drop-in OpenAI-compatible cache that replays repeated computer-use and agent workflows at near-zero cost.

Independent, source-attributed Hermes Agent workflows, skills and community projects. A static archive with public metrics and owner-reviewe

Turn any MCP server into a CLI in one command.

Search & read the login-walled Chinese internet for your AI agent — 微信公众号/小红书/抖音/B站 + Twitter/Reddit/HN/YouTube. One CLI, one JSON schema, z

Agent-native task management for humans and AI agents, built around a local-first API, SQLite storage, projects, calendar views, backups, an

Structured file edits for AI agents (JSON/YAML/TOML, markdown, AST, dry-run, MCP). Not a generic filesystem MCP.

Manage skills across Codex CLI, Claude Code, Cursor, OpenCode, and OpenClaw from one local app.

Safe MCP server for Plastic SCM / Unity Version Control (cm 10.0.16.6656+, UVCS 11.x). Readonly by default; guarded write tools.

Local-first MCP server for Obsidian vaults — no plugins, no API keys, with schema-validated Bases (.base) support.

Open-source browser automation broker for AI agents: persistent Chrome profiles, proxy-aware identities, remote API, MCP tools, browser-use/

Open MIT JSON Schema 2020-12 spec for SlimSnap captures. Bounding boxes, OCR text, extracted colors, annotations with intent and target_ref.

Download, extract and open Hélène. A desktop home for independent agents with lasting memory, tools, Telegram and optional local voice. Powe

Experience Microsoft Build, directly from your terminal, with GitHub Copilot CLI and the Microsoft Build CLI.

The context spine that 10x's every AI coding session. Live in 8 IDEs (Claude Code, Cursor, Cline, Continue, Aider, Codex, Windsurf, Zed) via

Durable MCP control plane for long-running Codex Desktop tasks

The marketplace for verifiable AI outcomes: state a task, get a fixed price upfront, and Primitive Bench routes across tools to complete it,

Web tasks that compile. Other browser tools tell your AI agent the click worked. chrome-agent reads the page back after every action and ans

The runtime reliability layer for AI agents.

Cross-agent plugin for manually excluding and editing AI conversation context, with search, filtering, hide/restore, undo, and original Sess

Agent-native Stata CLI for AI coding agents (Codex, Claude Code, Cursor). Run .do files, inspect data, and return econometric results as JSO

Open-source metasearch backend, MCP server, and AI search API for LLM agents. Python FastAPI search gateway with Google search via SerpBase

Tool-to-Agent Protocol: tools can be smart without embedded LLM calls.

面向 AI Coding Agent 的任务感知工作区上下文检索与生命周期追踪：按当前任务找到、组织并持续追踪最相关的文档、代码与媒体。纯本地、零模型调用、零网络。 Task-aware workspace context retrieval and lifecycle tra

Bounded, inspectable LLM inference pipelines from declared YAML — runs offline against Ollama or any OpenAI-compatible local server, emittin

oo is OOMOL's CLI toolkit. Everything can be done in the CLI.

local-first MCP discovery runtime for agents — search before invoke, reduce prompt bloat, and route to local MCP tools

Multi-account Google Workspace (Gmail, Drive, Calendar, Sheets, Docs) + Slack MCP servers for Claude Code. Setup guide with CLAUDE.md for AI

Hermes Agent plugin & skill for Twitter search, monitoring, follower exports & X actions. Not affiliated with X Corp.

Index, search, summarize, and approval-gate local Codex sessions through OpenClaw/MCP. Agent-first recall and dry-run orchestration for Code

让 Agent 操控 Windows 桌面软件

Verified control of the whole desktop for agents on Windows and Linux: apps with no API, windows, shell, browser. MCP server + CLI + REST AP

Go library, CLI, and MCP server for Camoufox browser automation.

Verdict-first output for AI coding agents. Tiny prompt + installer for Claude Code, Codex, Gemini, Cursor, opencode, and 30+ agents.

Gaia is a living, open registry of every AI agent skill — structured as a leveling skill graph. Live site: https://gaiaskilltree.com/

Maintain the same context between AI chats, models, and coding agents. Stop re-explaining your project every time you switch AI tools.

Local-first dashboards your AI agent builds and keeps maintaining, over data that never leaves your machine.

Lightweight file host and upload tools for coding agents to add artifacts to pull requests. Add CI checks for screenshots. Cloud and self-ho

Hand a whole browser task off in one call: a decision model drives the page server-side, so a flow costs one call, not a turn per click. Ref

Hermes Doctor для диагностики и approval-gated repair planning в Hermes Agent.

Agentizable: an open-source bounty network where AI agents post, fund, claim, solve, verify, and get paid for digital work.

Local-first, single-binary router for discovering and loading large agent skill libraries without flooding the model context.

Shared project knowledge and handoff for coding agents across long-running projects

Search your Claude Code conversation history by content (not just titles), then resume any session. grep + fzf, no index.

Agent tools from many descriptions in a single integration. Authorize operations, not just tool calls, and approve them when a human is requ