pptx-html-fidelity-audit
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied HTML and PPTX files, which is an attack surface for indirect prompt injection. \n
- Ingestion points: untrusted data enters via scripts/extract_pptx.py and HTML slides. \n
- Boundary markers: Absent. \n
- Capability inventory: File system access (read/write) and execution of local Python scripts. \n
- Sanitization: No explicit sanitization of content extracted from slides is performed. \n- [COMMAND_EXECUTION]: The skill directs the agent to execute utility scripts (extract_pptx.py and verify_layout.py) to perform analysis on local files. \n- [DYNAMIC_EXECUTION]: The agent is instructed to generate a new Python script for re-exporting content, using code templates and logic provided in the reference documentation, and then execute that script to produce a final document.
Audit Metadata