writing-codeceptjs-tests

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). This skill explicitly opens real web pages via the MCP server and ingests untrusted page content (ARIA snapshots and page HTML) during the "7. Open a live session via MCP" and "8. Learn the page and pick locators" workflow to choose locators and next actions, so third-party page content can influence subsequent tool use.