web-task

name	web-task
description	Web research and browser automation. Use when: scraping a page, filling out a form, taking a screenshot, navigating a dynamic JS app, doing provenance-tracked web research, or when context7 doesn't cover a web resource. Uses playwright MCP for dynamic pages and context-mode for static provenance-tracked fetches.
argument-hint	URL and task description (e.g. 'screenshot the login page at http://localhost:3000')

Web Tasks — Playwright and context-mode

Two tools, different use cases. Pick the right one or you're wasting resources.

Tool Selection

Scenario	Use
Dynamic JS app, login flows, forms, screenshots	playwright MCP
Static page, need citation/provenance trail	context-mode MCP
API docs / library reference	Use research-docs skill instead
Simple JSON API response	fetch / curl in terminal — don't spin up a browser

Playwright — Dynamic Pages

mcp_wulfnet-playw_playwright_get_tool_schema(tool_name)
mcp_wulfnet-playw_playwright_invoke_tool(tool_name, tool_input)

Always call get_tool_schema first.

Common Operations

Navigate and screenshot:

tool_name: "navigate"
tool_input: { "url": "<url>" }
// then
tool_name: "screenshot"
tool_input: { "name": "<descriptive-name>", "fullPage": true }

Click an element:

tool_name: "click"
tool_input: { "selector": "<css-selector or text>" }

Fill a form:

tool_name: "fill"
tool_input: { "selector": "<input selector>", "value": "<value>" }

CRITICAL: Always close the page when done:

tool_name: "close"

Procedure for UI Verification

1. Navigate to the target URL
2. Take a full-page screenshot to establish baseline
3. Perform the action (click, fill, submit)
4. Take another screenshot to confirm the result
5. Close the browser context
6. Report findings with screenshot references

context-mode — Provenance-Tracked Fetch

mcp_wulfnet-conte_context-mode_get_tool_schema(tool_name)
mcp_wulfnet-conte_context-mode_invoke_tool(tool_name, tool_input)

Use for static pages where you need a verifiable citation. Slower than playwright but produces a citation block suitable for reports.

Always call get_tool_schema first to confirm parameter names.

Rules

• Never use playwright for static content — it costs real CPU and a browser context
• Always close browser pages you open — leaked contexts accumulate
• Playwright screenshots go to the project's /screenshots/ directory if available
• If a page requires authentication, note what credentials are needed before starting
• Do not store credentials in code or screenshots — redact before saving