save-as-pdf
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads user-provided HTML design files to determine layout and content, which introduces a surface where malicious instructions could be embedded in the design data to influence agent behavior.
- Ingestion points: The skill reads the current HTML design file in Step 1 to understand its structure.
- Boundary markers: The instructions do not include the use of delimiters or 'ignore' instructions for the content being read.
- Capability inventory: The skill utilizes file reading, file writing, and platform tools such as show_html and show_pdf_export_dialog.
- Sanitization: There is no mention of sanitizing, escaping, or validating the HTML content before the agent processes it.
Audit Metadata