← Back to Skills Marketplace
173
Downloads
1
Stars
0
Active Installs
5
Versions
Install in OpenClaw
/install sensecraft-ai-model-hub
Description
Connect OpenClaw to the SenseCraft public model library to search, inspect, export, and download AI vision models for software and edge AI workflows.
Usage Guidance
This skill appears to be what it says, but take the usual precautions before running code that downloads and executes artifacts from the network: 1) Review the SenseCraft domain (https://sensecraft.seeed.cc) and confirm you trust it and the license of any downloaded models. 2) Run the demo in an isolated environment (VM or dedicated workstation) and use the provided virtualenv to contain Python packages. 3) Be cautious installing third-party PyPI packages (notably ai-edge-litert); consider reviewing package provenance or pinning versions. 4) Inspect downloaded model files before integrating them into production (models are binary artifacts and could trigger vulnerabilities in interpreters or custom ops). 5) Expect network and disk usage for bulk crawls/downloads; throttle or filter when necessary. 6) The demo accesses your webcam—ensure you understand and accept that and grant camera permission only to trusted terminals/apps.
Capability Analysis
Type: OpenClaw Skill
Name: sensecraft-ai-model-hub
Version: 1.1.2
The skill bundle provides a legitimate interface for searching, indexing, and downloading AI models from the Seeed Studio SenseCraft public model hub (sensecraft.seeed.cc). The Python scripts (sensecraft_models.py and sensecraft_webcam_person_demo.py) use standard libraries and reputable dependencies like ai-edge-litert for TFLite inference. The code includes proper input sanitization (safe_name function) to prevent path traversal during downloads, and the SKILL.md instructions are strictly aligned with the stated purpose of model discovery and local demonstration without any signs of prompt injection or malicious intent.
Capability Assessment
Purpose & Capability
The name/description match the included files and runtime instructions: Python scripts call SenseCraft public API endpoints, crawl/list/view model metadata, download model artifacts, and run a local webcam demo. There are no unrelated requested credentials, bizarre binaries, or config paths that don't belong to a model-hub + demo workflow.
Instruction Scope
SKILL.md and the scripts keep behavior within the stated scope: they call the documented SenseCraft endpoints at https://sensecraft.seeed.cc/aiserverapi, write artifacts under the skill root (./models, ./captures, .venv), and run a local demo that uses the webcam. The instructions do not attempt to read arbitrary system files, environment secrets, or other skills' configs. One expected aspect: the tool will download arbitrary model files from external URLs returned by the API (this is the point of the skill).
Install Mechanism
There is no platform-level install spec (instruction-only skill with bundled scripts). The included setup script creates a Python venv and installs PyPI packages (numpy, opencv-python, pillow, ai-edge-litert). Installing third-party Python packages is a normal part of this demo flow but does carry the usual supply-chain risk (unreviewed PyPI packages).
Credentials
The skill declares no required environment variables, no credentials, and no sensitive config paths. Scripts accept an optional PYTHON_BIN override but otherwise do not demand secrets or unrelated keys.
Persistence & Privilege
The skill is not always-enabled and is user-invocable. It does not request elevated or persistent system privileges and does not modify other skills or system-wide agent settings. Autonomous invocation by the agent is permitted (platform default) but is not combined with other concerning privileges here.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install sensecraft-ai-model-hub - After installation, invoke the skill by name or use
/sensecraft-ai-model-hub - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.1.2
Refined the store summary to a cleaner single-line product description for better search and listing display.
v1.1.1
Improved local demo usability with setup and launcher scripts, standard models/captures directories, and a macOS-friendly one-click runner.
v1.1.0
Added a proven local webcam inference demo for SenseCraft TFLite models, plus integration notes for macOS, ai-edge-litert, and SSCMA-Micro-based output decoding.
v1.0.1
Refined store summary wording for clearer public listing presentation.
v1.0.0
### v1.0.0
- Initial release of SenseCraft AI Model Hub
- Connects OpenClaw to the SenseCraft public model library
- Supports listing public models from SenseCraft
- Supports resolving model details and real download URLs by model ID
- Supports filtering by keyword, task, uniform type, and model format
- Supports exporting model indexes to JSON and CSV
- Supports downloading single or multiple model files
- Adds retry handling and safer bulk-fetch behavior
- Adds download manifest generation
- Adds metadata hints such as filename, extension, and TFLite-like URL detection
- Designed for both software-based and edge AI vision workflows
Metadata
Frequently Asked Questions
What is SenseCraft AI Model Hub?
Connect OpenClaw to the SenseCraft public model library to search, inspect, export, and download AI vision models for software and edge AI workflows. It is an AI Agent Skill for Claude Code / OpenClaw, with 173 downloads so far.
How do I install SenseCraft AI Model Hub?
Run "/install sensecraft-ai-model-hub" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is SenseCraft AI Model Hub free?
Yes, SenseCraft AI Model Hub is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does SenseCraft AI Model Hub support?
SenseCraft AI Model Hub is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created SenseCraft AI Model Hub?
It is built and maintained by Jancee (@jancee); the current version is v1.1.2.
More Skills