Senior Computer Vision
v2.1.1Computer vision engineering skill for object detection, image segmentation, and visual AI systems. Covers CNN and Vision Transformer architectures, YOLO/Fast...
⭐ 2· 2k·13 current·13 all-time
byAlireza Rezvani@alirezarezvani
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description (object detection, segmentation, optimization, deployment) align with the included SKILL.md and the three code scripts plus reference docs. The required capabilities (PyTorch, ONNX, TensorRT, etc.) are consistent with the stated purpose and show up in the guidance and code.
Instruction Scope
SKILL.md explicitly instructs the agent to run the included scripts and common CV tooling (yolo, train_net.py, export/benchmark commands). The runtime instructions operate on local datasets/models and produce config/export files as expected. Important security note: several scripts (inference_optimizer.py) call torch.load and other model-parsing libraries — loading untrusted model files with torch.load can execute arbitrary pickle payloads. The instructions do not tell the agent to read unrelated system files or to call external endpoints, so scope is appropriate, but exercise caution when pointing these tools at untrusted artifacts.
Install Mechanism
No install spec is present (instruction-only), and no remote downloads or archive extraction are specified. This reduces supply-chain risk because nothing is fetched/installed automatically by the skill bundle.
Credentials
The skill declares no environment variables, credentials, or config paths. The frameworks it uses (torch, onnx, tensorrt) are appropriate for its functions. There are no requests for unrelated secrets or system credentials.
Persistence & Privilege
Flags show normal (always: false, model invocation allowed). The skill does not attempt to persist itself or modify other skills or system-wide agent settings. It will read/write dataset and model files as part of normal operation (e.g., exports, calibration caches).
Assessment
This skill appears coherent for production CV work and contains helpful scripts and docs. Before running anything: (1) Inspect the scripts locally (they are included) and run them in a contained environment (VM, container, or isolated user) with a Python virtualenv. (2) Do not run analysis commands (e.g., inference_optimizer.py) on model files from untrusted sources — torch.load can execute code via pickle. (3) Be prepared for scripts to read/write files (dataset conversions, ONNX/TensorRT exports, calibration caches) and to require heavy dependencies (PyTorch, ONNX, TensorRT). (4) Install dependencies from trusted sources and limit network access if you want extra safety. If you need, I can point out the exact lines that call torch.load / deserialize operations and where files are written so you can audit them before running.Like a lobster shell, security has layers — review code before you run it.
latestvk97cja3pxs4y5g1wxzmrf8q7ph82jyey
License
MIT-0
Free to use, modify, and redistribute. No attribution required.