Agent-driven browser automation: a Chrome extension + Node bridge exposing composable browser_* tools for scanning, execution, and evidence capture.
Execute a native browser command (click, type, navigate, etc.) with optional verification.
Execute JavaScript in the current tab and optionally verify the result with an expect clause.
Observe the current page state with accessibility tree, visual capture, action space, and network events. Returns structured page observation including packets, task context, bundles, visual references, content, outline, action space, relations, causal events, tree diffs, and collections.
Retrieve the execution result, assertions, and evidence of a previously executed operation by operationId.
Capture a screenshot of the current tab.
List all tabs in the browser or navigate to a specific tab.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 14 | 2025-06-18+ | v2 |