4.3 NodeHub Case Study
NodeHub is D-Robotics' official case sharing platform. Unlike ClawHub, which shares single-file SKILL.md documents, NodeHub shares complete project code (including README, dependency declarations, and optionally a companion SKILL.md).
Division of Responsibilities Among Platforms
| Platform | Content | Use Case | Access |
|---|---|---|---|
| NodeHub | Complete project examples (with board type and category) | End-to-end examples ready to run out-of-the-box | https://developer.d-robotics.cc/en/nodehub |
| ClawHub | Single-file SKILL.md skills | Adding "operational strategies" for AI Agents | Within Studio: Skill Workshop → ClawHub Community |
| GitHub | Any code | General-purpose code hosting, without dedicated RDK categorization | github.com |
NodeHub is a specialized aggregation platform tailored for the RDK ecosystem, organized by board type and application category, enabling RDK developers to easily find reference implementations that "just work."
Publishing a Case Study to NodeHub
Step 1: Prepare Your Project
Ensure your project includes:
| Required | Recommended | Optional |
|---|---|---|
| Source code (public repository or ZIP archive) | README documentation (how to run it, dependencies) | SKILL.md (enables direct invocation by AI) |
| Supported board type description | Screenshots / GIFs demonstrating functionality | Video tutorial link |
Step 2: Log in to NodeHub
Visit https://developer.d-robotics.cc/en/nodehub and log in with your D-Robotics developer account.
Step 3: Create a Case Study
Click Publish Case, and fill in:
- Case title
- One-sentence description
- Target board(s) (X3 / X5 / S100 / Generic)
- Category tags (Vision, ROS, Control, AI Inference, etc.)
Step 4: Upload Content
| Method | Suitable For |
|---|---|
| Link to GitHub repository | Code already hosted on GitHub (recommended) |
| Upload ZIP package | Scenarios where Git isn't preferred |
Step 5: Review
After submission, your case will be reviewed by the NodeHub team. Once approved, it becomes publicly visible to all RDK developers.
Obtaining Case Studies from NodeHub
The NodeHub website offers the following browsing options:
| Browsing Method | Access Point |
|---|---|
| Browse featured recommendations | NodeHub homepage |
| Filter by board / category | Top navigation bar |
| Keyword search | Search box at the top |
| View details | Click case title → detail page (includes run instructions + user reviews) |
Ways to obtain the code:
| Method | Action |
|---|---|
| Direct git clone | Use the GitHub link provided on the case page |
| Convert to skill in Studio | Skill Workshop → Import from URL → Paste NodeHub case URL → AI automatically generates SKILL.md |
| Download ZIP directly | Click the Download button on the case page |
Currently, Studio does not support browsing NodeHub cases within Studio and one-click deployment to the board. Developers must manually download the code from the NodeHub website and then transfer it to the board using Studio’s File Manager or Remote Terminal.
Collaboration Between Cases and Skills
Many NodeHub case studies include a companion SKILL.md file. This “code + skill” combination allows a single case to serve two types of users:
- Developers:
git clonethe repository and refer to the implementation - AI Agents: Import the SKILL.md to learn “how to run this case”
Typical workflow:
- Find a ROS camera detection case on NodeHub
git cloneit to the board and manually run it once following the README to verify functionality- Import the companion SKILL.md into Studio via Skill Workshop → Import from URL
- Later, say in AI Dock: “Run that camera detection case,” and the Agent will execute it following the steps in SKILL.md
Characteristics of High-Quality Case Studies
| Feature | Description |
|---|---|
| Comprehensive README | Must include: environment requirements, installation steps, run commands, expected output |
| Multi-board compatibility | Uses conditional logic to handle board differences (e.g., if RDK_BOARD == 'X5') |
| Robust error handling | Doesn’t assume a perfect environment; provides helpful messages for common errors |
| Visuals (screenshots or video) | Instantly communicates “what this does” |
| Companion SKILL.md included | Enables direct use by AI Agents |
| Regular updates | Responds to RDK system upgrades and user feedback |
Before publishing, use these criteria to self-assess your project’s completeness.