data-report

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches frontend libraries from cdn.jsdelivr.net and cdn.tailwindcss.com. These are well-known and established services for hosting public web assets.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external data to generate its reports, which is a standard surface for indirect injection.
  • Ingestion points: Processes CSV, Excel, and JSON data provided by the user as specified in SKILL.md.
  • Boundary markers: No explicit instructions are provided to the agent to treat embedded text as data only.
  • Capability inventory: The skill utilizes fs:write and live-artifact to generate and display the resulting HTML report (defined in open-design.json).
  • Sanitization: The sample logic in example.html uses direct string interpolation into the DOM via innerHTML, which is an common pattern for report templates but represents a surface for data-carried content.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 03:22 AM
Security Audit — agent-trust-hub — data-report