Control persistent Chrome browser with automatic page capture.

Every DOM action (navigate, click, type, select, eval) auto-captures to the session dir:

• {prefix}.png — viewport screenshot

• {prefix}.md — page content as structured markdown

• {prefix}.html — full rendered DOM

• {prefix}-console.txt — browser console messages

Prefer reading these files to using 'extract' or 'screenshot' whenever possible. Actions on a page showing token-shaped content (Slack/GitHub/1Password tokens, otpauth:// URIs) or any data-sen-secret element write no capture files, token-shaped values are redacted from all output, and eval refuses outright while any data-sen-secret element is on the page — an accident guard against a cooperating agent's own reads, not a security boundary. A secret with no distinctive shape (bare TOTP seed, backup code) is captured like any other text until it is marked: mark it with 'set_attr' name 'data-sen-secret' as soon as the action that revealed it returns, before any other action on that page. Capture files written before the mark keep the value and are not deleted, so don't read them back. 'set_attr' is a write-only setter for exactly 'data-sen-nonce' or 'data-sen-secret' that is exempt from that eval restriction (see 'help' for details) — use it, not eval, to write onto the page while a secret is present.

Schema: 4 parameters — action, selector (CSS/XPath or null), payload (string or object), timeout (ms). selector targets a DOM element (null/omit for navigation, eval, tab management, etc.). payload is a string for simple actions (navigate=URL, type=text, eval=JS, keyboard_press=key). payload is an object for structured actions (set_viewport={width,height}, drag_drop={target}, etc.) — a JSON-encoded string of the same object works too. Tabs are tracked as sticky state; use switch_tab to change the active tab. Use action='help' for full per-action payload shapes.