Skills
Navigate
skills/mims-harvard/tooluniverse/setup-monocle3-remote-tool/Gen Agent Trust Hub

setup-monocle3-remote-tool

Pass

Audited by Gen Agent Trust Hub on Oct 11, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The installation instructions fetch the Monocle 3 R package from the official lab repository on GitHub (cole-trapnell-lab/monocle3) and a Python wheel for tuplatform-connect from connect.aiscientist.tools.
  • Evidence: remotes::install_github("cole-trapnell-lab/monocle3") in SKILL.md.
  • Evidence: pip install ... "tuplatform-connect @ https://connect.aiscientist.tools/downloads/..." in SKILL.md.
  • Note: The Python wheel is pinned with a SHA-256 hash to verify the integrity of the downloaded binary.
  • [INDIRECT_PROMPT_INJECTION]: The skill provides instructions to ingest and process biological data files (e.g., .h5ad) which can potentially contain embedded instructions if the data source is untrusted.
  • Ingestion points: The adata_path parameter in the run_monocle3_pseudotime tool invocation (documented in SKILL.md).
  • Boundary markers: No specific boundary markers or delimiters for the data content are defined in the tool invocation examples.
  • Capability inventory: The skill facilitates execution of complex R-based bioinformatics analysis via MCP tools.
  • Sanitization: The instructions recommend verifying input licenses and caution against using arbitrary caller-supplied model paths.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 11, 2026, 05:27 PM