i4h-workflow-dataset-mimic

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches the i4h-workflows project code from the Isaac for Healthcare GitHub organization to access the core mimicry and expansion utilities.
  • [COMMAND_EXECUTION]: Runs bash and Python commands to resolve dataset paths, verify file integrity, and perform the trajectory augmentation.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests robotics recording data from HDF5 files to perform mimicry. 1. Ingestion points: Reads local .hdf5 files defined in the workflow runs directory. 2. Boundary markers: Absent. 3. Capability inventory: Directory creation, symbolic linking, and bash/python execution. 4. Sanitization: Employs a Python script to validate the HDF5 file structure and ensure it contains valid episodes before processing begins.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 07:49 AM
Security Audit — agent-trust-hub — i4h-workflow-dataset-mimic