Mind Map HTML Output
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by processing untrusted user data and embedding it into a generated HTML file which executes in the user's browser context.
- Ingestion points: User-provided free text, markdown headings, or structured hierarchies are parsed to build the mind map data in
SKILL.md. - Boundary markers: The
assets/template.htmlfile uses placeholders like__MINDMAP_DATA__for direct interpolation without explicit boundary markers or instructions to the agent to escape the resulting JSON string. - Capability inventory: The skill delivers a standalone HTML file via
present_filesthat includes a toolbar and interactive mind map logic. - Sanitization: There is no evidence of sanitization or filtering of user-supplied topic text before it is inserted into the JavaScript section of the template.
Audit Metadata