foundationpose-pipeline

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides instructions to execute several Python scripts located in the product checkout, such as tools/bop_adapt/adapt.py, script/run_pipeline.py, and script/evaluate.py. These commands are run from the project root using a local virtual environment.- [DYNAMIC_EXECUTION]: The skill uses ctypes.CDLL('libcudart.so.13') in a Python snippet to verify that the required CUDA runtime library is correctly loaded in the user's environment before starting GPU-intensive tasks.- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external BOP datasets, which include images and JSON metadata.
  • Ingestion points: Dataset loading in adapt.py and run_pipeline.py.
  • Boundary markers: None identified.
  • Capability inventory: Local script execution and file writing to output directories.
  • Sanitization: Not explicitly described in the instruction text.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 01:08 PM
Security Audit — agent-trust-hub — foundationpose-pipeline