playwright-roll

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and inspects public upstream content—e.g., cloning/reading ~/code/playwright and walking docs/src/api/ commits and downloading the new driver (which provides playwright/driver/package/api.json from https://cdn.playwright.dev)—and the agent is instructed to read and act on that untrusted, user-generated documentation to decide ports/mismatches, so third-party content can materially influence actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill explicitly downloads and unpacks the Playwright driver at runtime from https://cdn.playwright.dev/builds/driver/playwright--linux.zip (via the wheel build), which provides remote executable code that the roll depends on and may be executed during testing, so it is a runtime external dependency that can execute remote code.