AI Toolbox is a cross-platform desktop application for managing the configuration, providers, models, prompts, plugins, sessions, MCP servers, and skills used by several AI coding assistants. Developers use it to switch and synchronize tools such as OpenCode, Claude Code, Codex, Gemini CLI, OpenClaw, and Pi across local, WSL, and remote SSH environments.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/coulsontl/ai-toolbox/agents-mdgit clone --depth 1 https://github.com/coulsontl/ai-toolboxWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/coulsontl/ai-toolbox/agents-md)<a href="https://agentmods.dev/instructions/coulsontl/ai-toolbox/agents-md"><img src="https://agentmods.dev/badge/instructions/coulsontl/ai-toolbox/agents-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.19144 | $0.19144 |
| Opus 5 | $0.09572 | $0.09572 |
| Sonnet 5 | $0.03829 | $0.03829 |
| Haiku 4.5 | $0.01914 | $0.01914 |
Grade A, and why
ai-toolbox AGENTS.md scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- 原因不是“风格统一”，而是实战兼容性：某些机器上 `reqwest + Schannel` 会在 TLS 建连阶段直接报 `SEC_E_NO_CREDENTIALS` / “安全包中没有可用的凭证”，表现为浏览器、Node、curl 正常，但 Rust `send()` 在真正发出业务请求前就失败。 How it starts
The opening of the file, as written. The whole thing — 1,064 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md - AI Toolbox Development Guide
This document provides essential information for AI coding agents working on this project.
Communication Language
与用户的所有对话必须使用中文，包括问题澄清、方案说明、进度反馈和结果总结。代码注释和 commit message 仍使用英文。
Module Documentation
主要功能模块可以在自身目录下放置 AGENTS.md，用于记录该模块的 Source of Truth、设计决策、关键流程、易错点和最小验证。修改模块代码前，如果目标目录或更近的父目录存在模块级 AGENTS.md，必须先阅读。
Hard Rules
- 修改任何模块目录下的文件前，若该模块目录或更近的父目录存在
AGENTS.md，必须先读，再设计或实现。 - 当一次改动引入新的非显而易见设计决策、修复未来高概率复发的问题、或重构跨文件关键流程时，应在同一任务内同步更新对应模块的
AGENTS.md。 - 根
AGENTS.md只保留全局规则。模块专属内容应下沉到对应模块目录下的AGENTS.md。 - 模块级
AGENTS.md只写高价值信息：Source of Truth、Why、关键流程、Gotchas、最小验证。不要写文件清单、类型定义、API 清单、数据库字段表、或可被代码直接证明的事实。 - 当根与模块
AGENTS.md冲突时，以作用域更近、语义更具体的文档为准。 - 新增模块级
AGENTS.md时，需在本节索引中同一任务内补上入口。 - 模块级
AGENTS.md建设处于迁移期时，根AGENTS.md中已有的高价值信息只能补充重组，不能因为下沉而删弱；确认模块文档已完整覆盖前，不要移除根中的对应规则。 - 后续开发中，只要遇到值得长期沉淀的知识、用户反复强调的约束、或已经复发/高风险复发的坑点，必须在同一任务内及时写回对应模块的
AGENTS.md，不能等“下次再整理”。 - 如果某条经验已经上升为跨模块通用规则、全局架构约束、或多个模块都会反复踩到的铁律，应同步写回根
AGENTS.md，而不是只埋在单个模块文档里。 - 数据库相关经验默认按全局规则处理，优先写回根
AGENTS.md；只有某条数据库约束严格局限于单一模块时，才在对应模块AGENTS.md补充模块特有语义。 - 仓库内用于给 agent 阅读的模块文档
AGENTS.md不是应用运行时资源；本地开发 watcher 和类似热重载链路应尽量忽略它们，避免把文档编辑误当成代码改动。 - 不要把上一条误用到产品运行时 prompt 文件上。当前仓库里的 OpenCode / Codex 运行时 prompt 文件名就是
AGENTS.md，Claude Code 运行时 prompt 文件名是CLAUDE.md；它们属于真实业务数据，备份、恢复、WSL/SSH 同步和页面交互都依赖这些文件，不能按“仅 agent 文档”排除。 - 修改
web/**中任何可见 UI、样式、布局、组件视觉、交互密度、空态、图标、颜色、字号、间距或弹窗表单前，必须先完整阅读根目录DESIGN.md。没有读取DESIGN.md就不得设计方案、不得写前端 UI 代码、不得声称遵循项目设计系统。
Template
docs/module-agents-template.md：模块级AGENTS.md模板。新增模块文档时先复制，再删除不适用小节。
Index
| 模块目录 | 说明 |
|---|---|
tauri/src/coding/ |
Coding 域共享规则：runtime location、事件驱动托盘、WSL Direct、跨工具 CLI/路径语义 |
tauri/src/coding/auth_refresh/ |
官方账号 OAuth 共享调度：启动首次 + 周期 ensure_fresh（Grok/Codex/Gemini） |
tauri/src/coding/claude_code/ |
Claude Code 后端配置、prompt、plugin、MCP 与 WSL 同步约束 |
tauri/src/coding/codex/ |
Codex 后端配置、auth/config.toml、prompt、plugin 与 WSL 同步约束 |
tauri/src/coding/claude_desktop/ |
Claude Desktop 3P profile 配置（deploymentMode+configLibrary+_meta）、Direct 应用、快照回滚与恢复官方；网关接管由 proxy_gateway 的 GatewayCliKey::ClaudeDesktop 驱动 |
tauri/src/coding/hermes/ |
Hermes Agent 运行时 config.yaml 的 custom_providers/模型/MCP 可视化与同步 |
tauri/src/coding/grok/ |
Grok CLI 后端 provider、config/auth、官方账号、prompt、plugin 与同步约束 |
tauri/src/coding/kimi/ |
Kimi Code CLI 后端 provider、config/credentials、官方账号、prompt 与同步约束 |
tauri/src/coding/gemini_cli/ |
Gemini CLI 后端配置、env/settings、prompt、usage、tray 与 WSL/SSH/备份同步约束 |
tauri/src/coding/mcp/ |
MCP Server 后端存储、工具配置同步、导入导出与 WSL 联动 |
tauri/src/coding/open_code/ |
OpenCode 后端配置文件、provider、prompt、tray 与 WSL 同步约束 |
tauri/src/coding/open_claw/ |
OpenClaw 后端配置文件与 WSL 同步约束 |
tauri/src/coding/oh_my_openagent/ |
Oh My OpenAgent 后端配置、临时本地态、应用链路与 OpenCode WSL 联动 |
tauri/src/coding/oh_my_opencode_slim/ |
Oh My OpenCode Slim 后端配置、临时本地态、应用链路与 OpenCode WSL 联动 |
tauri/src/coding/oh_my_pi/ |
Oh My Pi 运行时根目录、models.yml provider、config.yml 设置与本地 MCP/Skills 路径边界 |
tauri/src/coding/proxy_gateway/ |
本机代理网关、CLI 接管 manifest、配置备份恢复与模型级健康/日志文件 |
tauri/src/coding/proxy_gateway/transformer/ |
网关协议转换独立模块：Anthropic/OpenAI Chat/OpenAI Responses/Gemini Native JSON 与 SSE 互转 |
tauri/src/coding/session_manager/ |
会话浏览、详情、重命名、导入导出与运行时路径解析 |
tauri/src/coding/skills/ |
Skills 中央仓库、导入发现、同步、托盘和 WSL/SSH 相关链路 |
tauri/src/coding/tools/ |
Skills/MCP 共用工具适配、检测与自定义工具存储 |
tauri/src/coding/wsl/ |
WSL 同步配置、自动同步监听、WSL Direct 状态消费 |
tauri/src/coding/ssh/ |
SSH 连接、文件映射、手动同步、MCP/Skills 远端同步 |
tauri/resources/ |
编译期嵌入的模型默认数据资源：preset_models.json/models.dev.json 的来源、顺序语义与缓存边界 |
web/features/coding/claudecode/ |
Claude Code 前端页面、根目录配置、provider 与 prompt 交互 |
web/features/coding/claudedesktop/ |
Claude Desktop 前端页面、provider/模型映射与通用配置交互（复用 claudecode 样式与网关接管按钮） |
web/features/coding/hermes/ |
Hermes 前端页面、custom_providers 与模型设置交互（复用 pi 样式） |
web/features/coding/codex/ |
Codex 前端页面、根目录配置、provider 与 prompt 交互 |
web/features/coding/grok/ |
Grok CLI 前端页面、根目录配置、provider、官方账号、plugin、prompt 与 session 交互 |
web/features/coding/geminicli/ |
Gemini CLI 前端页面、根目录配置、provider、prompt、usage 与 session 交互 |
web/features/coding/kimi/ |
Kimi Code CLI 前端页面、根目录配置、provider、官方账号、prompt 与通用配置交互 |
web/features/coding/gateway/ |
Gateway 前端页面、统计/请求/设置 Tab、顶部入口与 visibleTabs 可见性 |
web/features/coding/image/ |
Image 前端页面、工作台、渠道管理、历史与结果交互 |
web/features/coding/mcp/ |
MCP 前端页面、服务器管理、导入流程与工具同步交互 |
web/features/coding/opencode/ |
OpenCode 前端页面、配置路径、provider、prompt 与模型刷新交互 |
web/features/coding/openclaw/ |
OpenClaw 前端页面、配置路径、provider 与配置文件交互 |
web/features/coding/shared/ |
coding 共享前端语义：根目录弹窗、全局 prompt、favorite provider、会话面板、连通性测试 |
web/features/coding/skills/ |
Skills 前端页面、中央仓库视角、分组展示与批量同步交互 |
web/features/settings/ |
WSL/SSH 设置页、同步入口、moduleStatuses 消费和 UI 边界 |
web/components/common/ |
共享编辑器与基础交互组件的性能和正确性约束 |
tauri/src/settings/backup/ |
备份恢复、WebDAV、自动备份与恢复后续链路 |
tauri/src/coding/image/ |
Image 后端渠道配置、任务、资产落盘、图片 API 调用与备份联动 |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- today Changed · +11 lines · +438 tokens per session e98999cbdb40
- yesterday Changed · +5 lines · +1,102 tokens per session 2ba8d7deaa13
- 5d ago First seen · 1,048 lines · 17,604 tokens per session scan A 0bbd9f2763e7
ai-toolbox AGENTS.md is an instructions file published in the GitHub repository coulsontl/ai-toolbox (1,363 stars, last pushed today), licensed MIT. It adds 19,144 tokens to every session, about $0.0957 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
deepseek-harness AGENTS.md
AGENTS.md instructions for deepseek-ai/deepseek-harness, covering agents.md, pre-stable apis and released session data, repository layout, commands and host sandbox failures.