The Agent Skills Directory

[DATA_EXFILTRATION]: The skill utilizes the upload_html_report tool to transmit browser session evidence, including videos, traces, and HTML reports, to Shiplight's cloud infrastructure. This functionality is intended for creating shareable verification results and is associated with the skill author's own services.
[PROMPT_INJECTION]: The skill reads and processes external web content, creating a surface for indirect prompt injection. 1. Ingestion points: DOM trees and screenshots via inspect_page, and browser console logs via get_browser_console_logs. 2. Boundary markers: No delimiters or instructions to ignore embedded commands are present in the skill's logic for processing retrieved content. 3. Capability inventory: The agent has the ability to interact with the browser (act), execute shell commands for server setup, and upload data to the cloud. 4. Sanitization: No sanitization, validation, or escaping of the ingested web content is described.
[COMMAND_EXECUTION]: The skill instructions direct the agent to execute local shell commands such as npm run dev or yarn dev to ensure the application's development server is running during the verification process.

verify