sd-parse-results

Pass

Audited by Gen Agent Trust Hub on Mar 5, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it extracts and processes untrusted data from external web pages.
  • Ingestion points: ScienceDirect search results (titles, authors, and journal names) are extracted directly from the DOM using document.querySelectorAll('li.ResultItem') within the Javascript block in SKILL.md.
  • Boundary markers: The skill does not implement boundary markers or 'ignore embedded instructions' warnings when returning the extracted strings to the agent context.
  • Capability inventory: The skill uses the evaluate_script tool for data extraction; while this specific skill lacks write or network capabilities, the data it returns could be used to influence downstream agent behavior.
  • Sanitization: The script employs .textContent, which effectively prevents HTML/script injection within the browser, but it does not perform semantic sanitization to filter out potential instructions directed at the AI agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 5, 2026, 06:45 AM