setup-scvelo-remote-tool

Pass

Audited by Gen Agent Trust Hub on Oct 11, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches installation requirements and environment configurations from 'connect.aiscientist.tools' and the author's GitHub repository to set up the ToolUniverse environment.
  • [EXTERNAL_DOWNLOADS]: Downloads and installs the 'tuplatform-connect' Python package from a remote URL. The installation command uses a SHA-256 hash to ensure the integrity of the downloaded wheel file.
  • [COMMAND_EXECUTION]: Instructs the user to execute various CLI tools including 'uv', 'pip', and the 'tu' utility to manage virtual environments, install packages, and serve the remote tool endpoint.
  • [INDIRECT_PROMPT_INJECTION]: The 'run_scvelo_velocity' tool processes external data files (e.g., '.h5ad'). This represents an ingestion surface where maliciously formatted data could attempt to influence the agent's behavior.
  • Ingestion points: 'adata_path' parameter in 'run_scvelo_velocity' tool.
  • Boundary markers: None specified in the prompt template.
  • Capability inventory: Local file access, Python script execution, and network connectivity via the 'tu serve' command.
  • Sanitization: The instructions advise verifying data rights and avoiding arbitrary caller model paths.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 11, 2026, 05:27 PM