Skills
skills/sungkhum/agent-skills/idml/Gen Agent Trust Hub

idml

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • PROMPT_INJECTION (LOW): The skill extracts text from untrusted IDML story files via extract_story_text.py, creating a surface for indirect prompt injection. Ingestion points: Stories/*.xml. Boundary markers: None. Capability inventory: File system write access (pack_idml.py) and local script execution via runpy (smoke_test.py). Sanitization: No natural language sanitization is performed.
  • COMMAND_EXECUTION (LOW): The smoke_test.py script utilizes runpy.run_path to execute local helper scripts within the same package. This is used for round-trip validation and is generally safe but qualifies as dynamic execution of local code.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:28 PM