page-cro

Pass

Audited by Gen Agent Trust Hub on Mar 5, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill is designed to analyze external web pages and marketing content provided by users, which introduces a potential surface for indirect prompt injection. * Ingestion points: User-provided URLs and marketing page content as described in SKILL.md. * Boundary markers: No delimiters or ignore embedded instructions warnings are present to isolate analyzed content. * Capability inventory: The skill contains no scripts or tools; it is purely informational markdown. * Sanitization: No sanitization or validation of the input content is performed.
  • [NO_CODE]: The skill is composed entirely of Markdown documentation and evaluation JSON files, with no executable code or scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 5, 2026, 11:10 PM