skills/cookjohn/sd-skills/sd-download/Gen Agent Trust Hub

sd-download

Pass

Audited by Gen Agent Trust Hub on Sep 30, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external ScienceDirect article and search result pages through browser automation tools.
  • Ingestion points: Browser evaluation scripts read the DOM of article pages and search results in SKILL.md.
  • Boundary markers: No specific boundary markers or instructions to ignore embedded content are used when processing page data.
  • Capability inventory: The skill utilizes navigate_page, evaluate_script, take_snapshot, and click to interact with the browser environment.
  • Sanitization: No explicit sanitization of the extracted page content is performed before processing.
  • [DYNAMIC_EXECUTION]: The skill employs evaluate_script to execute JavaScript logic within the browser session for DOM manipulation and state checking.
  • The skill uses an initScript during navigation to override the navigator.webdriver property, which is a common technique to prevent automated browser detection.
  • JavaScript is used to dynamically create and click hidden anchor tags to trigger file downloads to the local disk.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 30, 2026, 01:54 PM