Vouch request — implement web_scrape tool (#5052)

[priyansh19]

Hi maintainers 👋

I'd like to contribute an implementation of #5052 (Add tool: web_scrape) and am requesting a vouch so I can open the PR (per CONTRIBUTING).

What I've done so far: cloned the repo and traced the tool architecture. Tools are classes implementing AgentTool<schema, Details> in packages/coding-agent/src/tools/, registered in the factory map in tools/index.ts and listed in tools/builtin-names.ts, and gated by a settings.get(...) flag (mirroring the existing web_search tool).

Proposed approach for web_scrape:

• New WebScrapeTool in packages/coding-agent/src/tools/web-scrape.ts, modeled on WebSearchTool.
• Register it in tools/index.ts (web_scrape: s => new WebScrapeTool(s)) and add it to builtin-names.ts.
• Backends: Firecrawl and Kagi, selected by which API key is configured; return the page as Markdown (default) or raw/original. Emit a clean "backend not configured" error instead of failing silently.
• Settings gate web_scrape.enabled + backend key settings, following the web_search pattern.
• Prompt doc under src/prompts/tools/, focused tests, and a changelog entry.

This targets the issue's pain point — the read tool and a curl fallback both failing on JS-heavy or access-blocking pages — without the heavy Firecrawl CLI file-dump flow.

Could a maintainer !vouch @priyansh19 so I can open the PR? Happy to adjust scope or backend choices first. Thanks!