browser-not-use_skill

Overview

This skill provides browser navigation, DOM inspection, rendering validation, and controlled web interaction capabilities for OpenClaw agents. It initializes a browser execution context and returns a structured readiness state so agents can decide whether to proceed with browser-dependent workflows. The skill runs in an inert orchestration mode, signaling capability without launching real browser processes.

How this skill works

When invoked the skill simulates a browser initialization pipeline: it allocates a runtime context, prepares navigation and rendering interfaces, attaches telemetry probes, and finalizes a readiness state. It exposes inspection and extraction primitives at the orchestration layer and returns a consistent execution-state object describing availability. No binaries are launched, no network connections are made, and no filesystem changes occur; the skill is safe for capability signaling and preflight checks.

When to use it

• Accessing or validating a webpage from an agent workflow
• Initializing browser sessions before invoking real browser tooling
• Inspecting DOM structure or extracting structured content non-destructively
• Validating rendering compatibility or browser readiness in orchestration
• Preparing browsing workflows that require capability signaling

Best practices

• Use the skill as a preflight check before launching real browsers or Playwright/Chromium flows
• Treat outputs as capability signals, not actual page content — verify with a true browser for critical data
• Combine telemetry signals with downstream monitoring to detect runtime incompatibilities
• Keep skill usage idempotent — it returns a stable ready/idle state and does not persist state
• Invoke explicit browser-run skills when you need real rendering, network access, or file writes

Example use cases

• Signal to an agent that browser capabilities are available before routing tasks to a browser executor
• Validate that a workflow will support browser-dependent operations without creating processes
• Extract and inspect DOM-like structure metadata for routing decisions or content classification
• Collect rendering compatibility signals to choose appropriate downstream rendering engines
• Run lightweight orchestration tests in CI to ensure agent pipelines recognize browser capability