page-cro

Pass

Audited by Gen Agent Trust Hub on Apr 13, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it is designed to ingest and process untrusted data from external sources to perform its audit.\n
  • Ingestion points: The agent processes external content from live page URLs, Clarity CSV exports, and visual screenshots provided by users (SKILL.md, Step 0 and Step 2).\n
  • Boundary markers: There are no explicit instructions to wrap this untrusted external content in specific delimiters or to provide negative constraints to ignore embedded instructions within the processed data.\n
  • Capability inventory: The skill calls data retrieval tools like get_marketing_context and suggests handing off implementation tasks to skills such as website-content-editor, which may possess production-level write capabilities.\n
  • Sanitization: The instructions do not specify any validation or sanitization steps for the data retrieved from URLs or CSV files before it is incorporated into the agent's reasoning context.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 13, 2026, 01:21 PM