cnki-navigate-pages

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface where user-controlled arguments are interpolated into JavaScript templates for browser execution.
  • Ingestion points: The $ARGUMENTS provided by the user are used to replace ACTION_HERE and SORT_HERE placeholders in the script.
  • Boundary markers: No delimiters or specific instructions are included to isolate these arguments from the rest of the script logic.
  • Capability inventory: The skill utilizes evaluate_script, which allows it to interact with the DOM and perform actions in the browser context.
  • Sanitization: The logic includes basic sanitization, such as using regex to extract digits for pagination and a predefined map for valid sort options.
  • [COMMAND_EXECUTION]: The skill performs dynamic script generation by constructing JavaScript functions from templates and executing them using the evaluate_script tool.
  • Evidence: The skill documentation describes how to build and execute asynchronous JavaScript functions by injecting user-defined actions and sort parameters.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 04:58 PM