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/openlegged/urdf-studio/claude-mdgit clone --depth 1 https://github.com/OpenLegged/URDF-StudioWhat 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 | $0.08074 | $0.08074 |
| Opus 5 | $0.04037 | $0.04037 |
| Sonnet 5 | $0.01615 | $0.01615 |
| Haiku 4.5 | $0.00807 | $0.00807 |
Grade B, and why
URDF-Studio CLAUDE.md scanned grade B 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 2d ago.
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.
Asks for rootmediumPrivilege escalation
A mod that escalates privileges can change anything on the machine, not only the project.
sudo apt install -y mkcert libnss3-tools # Ubuntu/Debian；或下载二进制到 ~/.local/bin How it starts
The opening of the file, as written. The whole thing — 297 lines — stays where its author put it; the contents beside it link to each section on GitHub.
URDF Studio Agent Guide
最后更新：2026-08-04 | 技术栈：React 19.2 + TypeScript 5.8 + Three.js/R3F + Vite 6.2 + Tailwind CSS 4.1 + Zustand 5 完整文档索引：docs/CATALOG.md
URDF Studio 是机器人设计、装配、可视化与导出工作台。核心能力：单模式 Editor 编辑、多 URDF 组装与桥接关节、多格式导入导出（URDF / MJCF / SDF / USD / Xacro / ZIP / .usp）、AI 生成与审阅、PDF/CSV 报告、可复用 react-robot-canvas 画布封装。
语言偏好
- 默认中文沟通（过程说明、总结、澄清、错误说明），除非用户明确要求其他语言
- 代码标识符、命令、文件路径、API 名称、错误原文保持原文
需求澄清与第一性原理
- 使用第一性原理思考：从用户的原始需求、问题动机、目标结果和约束出发，不要只按表面措辞机械执行
- 不要默认用户已经完全清楚自己想要什么、为什么要这样做、以及应该怎样得到结果；需要主动识别目标、动机、范围和成功标准是否清晰
- 如果需求意图不清、目标互相冲突、实现路径会显著影响产品形态或可能引入错误抽象，先停下来用中文和用户讨论关键问题
- 如果任务目标明确且风险可控，继续自治执行；不要把可由代码、文档或上下文推断的问题变成无谓确认
src/ 目录结构
src/
├── app/ 应用编排层：App shell、viewer 组合、导入导出、canonical source documents、USD hydration
│ ├── components/ App 级 UI 编排与跨域入口（header/settings/snapshot-preview/unified-viewer/workspace/ai/export/snapshot-dialog）
│ ├── hooks/ 跨 store / viewer / source document hook（file-export/workspace-mutations/workspace-source-sync）
│ ├── utils/ App 层辅助逻辑与导入准备 wrapper（import-preparation/）
│ └── workers/ App 编排使用的 worker
├── features/ 业务功能模块
│ ├── ai-assistant/ AI 生成与审阅
│ ├── assembly/ 桥接组件创建与组装（桥接弹窗内部模块在 components/bridge-create/）
│ ├── code-editor/ 源码编辑器
│ ├── editor/ Editor 统一公开入口
│ ├── file-io/ 底层文件能力（格式检测 workflow wrapper、project archive、USD/SDF export、弹层）
│ ├── hardware-config/ 硬件/电机配置（兼容层 re-export）
│ ├── property-editor/ 属性编辑、几何编辑、碰撞优化
│ ├── robot-tree/ 文件树与结构树（tree-editor/ + tree-node/）
│ └── urdf-viewer/ Editor 实现：拓扑/几何/碰撞/测量 + renderer backend + USD runtime + workers
├── store/ Zustand 状态层
│ └── workspace/ workspaceStore 内部模块：bridge/component/topology CRUD、runtime、helpers、types
├── shared/ 共享组件、3D 基础设施、hooks、i18n、数据、调试桥接、utils、workers
│ ├── components/ 共享 UI（ui/）、3D 基础组件（3d/）、DraggableWindow、Panel、reports、纯 mesh renderer 组件
│ ├── data/ 静态数据与内置配置
│ ├── debug/ 调试桥接与验证辅助
│ ├── hooks/ 通用 React hooks
│ ├── i18n/ 本地化文案与类型
│ ├── utils/ 通用工具（assembly/ pdf/ reports/ robot/ three/ 等）
│ └── workers/ 共享 worker
├── core/ 纯逻辑：解析器、robot core、mesh loaders、parse workers、runtime diagnostics
│ ├── geometry/ 几何计算
│ ├── image-compressor/ 图像压缩逻辑
│ ├── loaders/ mesh / 资源加载
│ ├── parsers/ URDF / MJCF / SDF / USD 解析与格式检测
│ ├── robot/ robot core、组装 seed、runtime patch diff 与拓扑逻辑
│ ├── stl-compressor/ STL 压缩逻辑
│ ├── utils/ core 层工具
│ └── workers/ parse / runtime workers
├── lib/ 对外复用的 RobotCanvas 封装（仅收稳定通用能力）
├── styles/ 全局样式与语义 token
└── types/ 跨模块类型定义
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.
- 2d ago First seen · 297 lines · 8,074 tokens per session scan B 808ef44fe911
URDF-Studio CLAUDE.md is an instructions file published in the GitHub repository OpenLegged/URDF-Studio (467 stars, last pushed 2d ago), licensed Apache-2.0. It adds 8,074 tokens to every session, about $0.0404 per session on Opus 5. A static security scan graded it B with 1 finding (asks for root). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
bagel AGENTS.md
Instructions for Extelligence-ai/bagel, covering bagel for agents, set up bagel for a user, use bagel well and develop on bagel.
agenticros AGENTS.md
Instructions for agenticros/agenticros, covering agenticros, what this is, important: controlling the robot from codex, architecture and key source files.
sint-protocol CLAUDE.md
Instructions for sint-ai/sint-protocol, covering sint protocol — agent guide, what is this?, quick commands, monorepo layout and architecture rules.
KernelWiki CLAUDE.md
Instructions for mit-han-lab/KernelWiki, covering blackwell kernel optimization knowledge base — schema, navigation, recommended: use the skill scripts, direct navigation (when skill not available) and three-layer architecture.
autonomous-os AGENTS.md
Instructions for autonomous-ai/autonomous-os, covering agents.md, multi-ide rules, working style, parallel work / subagents and device access rules.
tilelang-ascend AGENTS.md
Instructions for tile-ai/tilelang-ascend, covering agents.md, 项目概述, 核心功能, skills 索引 and 算子开发编排体系（opencode 多代理）.