ai-for-science-diffsbdd
Pass
Audited by Gen Agent Trust Hub on Mar 28, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches project resources and model weights from established public repositories.\n
- Clones the DiffSBDD research project from GitHub.\n
- Downloads pre-trained model weights from Zenodo (zenodo.org), a well-known research data hosting service.\n- [COMMAND_EXECUTION]: Executes system commands for environment configuration and library compilation.\n
- Compiles the
torch_scatterlibrary from source usingpython setup.pyto ensure NPU compatibility.\n - Configures environment variables and installs dependencies via standard package managers (pip and conda).\n- [SAFE]: Includes a validation script for dependency and NPU runtime checking.\n
- The
scripts/validate_diffsbdd_env.pyscript verifies module availability and NPU backend readiness using restricted dynamic imports.\n - No suspicious or obfuscated behavior was detected in the skill instructions or supporting scripts.
Audit Metadata