skills/zaoqu-liu/scienceclaw/datamol/Gen Agent Trust Hub

datamol

Pass

Audited by Gen Agent Trust Hub on Apr 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a comprehensive guide for the datamol library, which is a well-known and legitimate open-source tool in the cheminformatics domain. No obfuscation or malicious logic was detected.\n- [COMMAND_EXECUTION]: Provides standard instructions for library installation via uv pip install datamol. This is normal for a developer-centric skill.\n- [EXTERNAL_DOWNLOADS]: Includes functionality to read datasets from remote URLs and cloud storage (S3, GCS, HTTP) using the fsspec library, which is a standard feature for large-scale data analysis.\n- [DATA_EXFILTRATION]: Documents the capability to write molecular data to remote cloud storage buckets, allowing for legitimate data export in computational chemistry workflows.\n- [PROMPT_INJECTION]: The skill identifies a surface for indirect prompt injection as it ingests external data (SMILES, SDF, CSV, Excel). However, this is the primary function of the tool and no exploitation was observed.\n
  • Ingestion points: Functions like dm.read_sdf, dm.read_csv, and dm.open_df.\n
  • Boundary markers: The instructions do not define prompt delimiters or warnings for embedded instructions.\n
  • Capability inventory: Includes file system read/write, network operations via fsspec, and multi-core parallel processing.\n
  • Sanitization: Provides structural and chemical sanitization (dm.sanitize_mol), though not specific text-based safety sanitization.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 16, 2026, 08:25 AM