cnki-journal-search

Warn

Audited by Gen Agent Trust Hub on Jul 17, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill's instructions in SKILL.md tell the agent to replace a placeholder QUERY_HERE inside a JavaScript string literal with user-supplied search terms. This creates a template injection vulnerability where a specially crafted input can escape the string and execute arbitrary JavaScript code within the browser context.
  • [PROMPT_INJECTION]: The skill processes untrusted data from the CNKI website, creating an indirect prompt injection surface.
  • Ingestion points: The skill extracts journal titles and metadata from the page using document.body.innerText and document.querySelectorAll in the Step 2 script.
  • Boundary markers: There are no delimiters or instructions to the agent to treat this external content as data rather than instructions.
  • Capability inventory: The skill uses the browser tool for navigation and script execution.
  • Sanitization: The skill does not perform any sanitization or validation on the data scraped from the website before returning it to the agent environment.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 17, 2026, 12:19 AM
Security Audit — agent-trust-hub — cnki-journal-search