The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill provides an eval command and a raw command that allow for the execution of arbitrary JavaScript and Chrome DevTools Protocol methods within the browser context.
Evidence: Commands node scripts/cdp.mjs eval "document.title" and node scripts/cdp.mjs raw Runtime.evaluate '{"expression": "navigator.userAgent"}' are documented as core features.
[COMMAND_EXECUTION]: The skill relies on executing local scripts (scripts/cdp.mjs) to interact with the Chrome browser daemon.
Evidence: Multiple CLI commands are documented to run via node scripts/cdp.mjs.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from external websites (HTML, text content, accessibility trees, and console logs) while possessing powerful write capabilities like form filling and JavaScript execution.
Ingestion points: snapshot, get-text, get-console, and navigate commands retrieve data from the active browser tab.
Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands in the retrieved web content.
Capability inventory: The skill can fill forms (fill), click elements (click), execute JavaScript (eval), and navigate to new URLs.
Sanitization: No explicit sanitization or filtering of the retrieved web content is mentioned before it is passed to the agent context.

faster-chrome-devtools-skill