i4h-workflow-dataset-teleop

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill clones the official i4h-workflows repository from GitHub (isaac-for-healthcare org) to provide the underlying robotics simulation and recording logic.
  • [COMMAND_EXECUTION]: The skill executes multiple bash scripts including run.sh for starting the simulation and stop.sh for lifecycle management, which are required for teleoperation recording.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and parses stdout logs from the teleoperation process to extract keybinding information for the user.
  • Ingestion points: The teleop.log file within the dynamically created run directory.
  • Boundary markers: Absent; the skill uses grep with regex patterns like 'Keybind' and 'Controls' to identify relevant data.
  • Capability inventory: The skill utilizes repository cloning, shell command execution, and file system modifications (mkdir, symlinks).
  • Sanitization: Absent; the skill extracts blocks of log data directly to report them to the agent context.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 07:50 AM
Security Audit — agent-trust-hub — i4h-workflow-dataset-teleop