NYC

huashu-slides

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: LOW
Full Analysis
  • [COMMAND_EXECUTION] (SAFE): The create_slides.py script performs image and slide processing using dedicated high-level libraries. It does not invoke any shell commands or utilize subprocess/os.system calls.
  • [EXTERNAL_DOWNLOADS] (SAFE): No remote scripts, binaries, or data are downloaded during execution. All dependencies are standard Python packages declared in the script metadata.
  • [DATA_EXFILTRATION] (SAFE): The skill does not perform network operations and lacks the capability to send data to external servers. File access is limited to reading user-provided images and writing the generated PPTX file.
  • [PROMPT_INJECTION] (SAFE): The reference materials include stylistic guidance for image generation but do not contain instructions aimed at subverting agent behavior or bypassing safety protocols.
  • [INDIRECT_PROMPT_INJECTION] (LOW): The script processes external image titles and PPTX templates. 1. Ingestion: args.titles, args.template. 2. Boundary markers: None. 3. Capability inventory: File read and write via python-pptx. 4. Sanitization: None. While an ingestion surface exists, the risk is LOW as the script treats input purely as data for inclusion in a non-executable document format.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 17, 2026, 04:41 AM