screencli-record

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

• Third-party content exposure detected (high risk: 1.00). The skill's docs (SKILL.md and references/cli-reference.md) require the agent to navigate and interact with an arbitrary target URL supplied by the user (e.g., "The AI agent navigates the target URL" and the required "npx screencli record [url] -p ''"), meaning it fetches and interprets untrusted public web content which can directly influence the agent's actions.