Skills
Navigate
skills/nexu-io/open-design/open-design-landing-deck/Gen Agent Trust Hub

open-design-landing-deck

Pass

Audited by Gen Agent Trust Hub on Sep 28, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The generator script scripts/compose.ts interpolates content from inputs.json into the generated HTML slides without sanitization.
  • Ingestion points: Slide text fields (title, body, bullets, quotes) in inputs.json which are populated based on user input or agent generation.
  • Boundary markers: No delimiters or warnings are used to signal the boundary between template code and user-provided text.
  • Capability inventory: The skill utilizes file-write and node-runtime capabilities to generate and save the HTML artifact.
  • Sanitization: Content is concatenated into HTML strings in scripts/compose.ts without escaping HTML special characters, creating an XSS vulnerability surface in the resulting deck.
  • [COMMAND_EXECUTION]: The skill provides scripts for composing the deck using npx tsx, which involves shell command execution to build the final artifact.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 28, 2026, 03:36 AM