datamol

Pass

Audited by Gen Agent Trust Hub on Nov 6, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs users to install the datamol package from a standard package registry using uv pip install.
  • [EXTERNAL_DOWNLOADS]: The documented library supports data ingestion from remote URLs and cloud storage providers including AWS S3, Google Cloud Storage, Azure, and via HTTP/HTTPS protocols using fsspec integration.
  • [COMMAND_EXECUTION]: The skill provides shell commands for package installation and utilizes parallel processing capabilities via the n_jobs parameter for batch molecular operations.
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves processing external molecular data files (SDF, SMILES, CSV, Excel) from potentially untrusted sources.
  • Ingestion points: Molecular file reading functions such as dm.read_sdf(), dm.read_csv(), and the universal dm.open_df() reader.
  • Boundary markers: Code snippets do not explicitly include boundary markers for data ingestion, though this is typical for data science documentation.
  • Capability inventory: Includes file writing (dm.to_sdf()), network access for remote files, and multi-core CPU utilization for descriptors and clustering.
  • Sanitization: The skill provides built-in functions for molecular standardization and sanitization (dm.standardize_mol(), dm.sanitize_mol()) to ensure chemical validity of ingested data.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 6, 2026, 12:15 PM