feat(client): add batch_fetch for concurrent multi-URL fetching

[chaliy]

What

Add batch_fetch() and batch_fetch_with_options() for fetching multiple URLs concurrently.

Why

Agents frequently need to fetch 5-10 pages to research a topic. Batch fetching reduces round-trips and lets the tool manage concurrency.

How

• batch_fetch(requests, concurrency) — simple API with default options
• batch_fetch_with_options(requests, options, concurrency) — full control
• Uses futures::stream::buffer_unordered for concurrent execution
• Default concurrency: 5, configurable via parameter
• Results returned in same order as input requests
• Each URL processed independently (one failure doesn't affect others)

Risk

• Low — new functions, no changes to existing behavior
• Uses existing fetch infrastructure

Checklist

• Unit tests (multi-URL, partial failure, concurrency, empty input)
• Clippy clean
• Docs build clean

Closes #75