email-newsletter-builder
Pass
Audited by Gen Agent Trust Hub on Nov 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest user-supplied content (text, titles, links, and image URLs) to populate an HTML newsletter template, which constitutes an indirect injection surface.
- Ingestion points: User input is gathered and stored in a
config.jsonfile which is then parsed by thescripts/generate_newsletter.pyscript (specifically thebody,excerpt, andtextfields). - Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are provided in the workflow.
- Capability inventory: The associated script performs local file reads and writes (
config.jsonandnewsletter.html). It does not have network access or the ability to execute arbitrary shell commands. - Sanitization: The Python script employs
html.escape()for critical fields like page titles, preheader text, and CSS color values. It intentionally allows raw content in text body fields to support legitimate HTML formatting (e.g., bolding, italics, links) as specified in the documentation. This is standard behavior for a template generation tool and poses minimal risk as email clients inherently restrict active content like JavaScript.
Audit Metadata