roboflow-custom-weights-upload
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is authored by Roboflow and describes the official workflow for interacting with their services. It includes defensive security instructions for the AI agent.
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to install necessary Python packages (
roboflow,rfdetr,python-dotenv,torch,ultralytics) using standard package managers. These are legitimate dependencies for the stated purpose of model weight packaging and upload. - [COMMAND_EXECUTION]: The skill involves executing shell commands and Python scripts to perform the upload. This is the primary intended function of the skill (a development tool) and is accompanied by preflight checks and user confirmation prompts.
- [CREDENTIALS_SAFE]: The skill demonstrates high security awareness regarding API keys. It explicitly instructs the agent NEVER to ask the user to paste keys into chat and NEVER to scan sensitive system files. It recommends using scoped keys via an MCP tool and managing them through
.envfiles with appropriate warnings against unsafe shell practices like sourcing untrusted files.
Audit Metadata