Add Parallel Search to the MCP store

[georgeatparallel]

Why

Adds an optional web search and URL-fetching MCP server to Peezy's existing MCP store. The catalog description makes clear that selected search queries and requested URLs are sent to Parallel.

Changes

• Add parallel-search beside the existing search and crawling entries.
• Connect Peezy's stdio-only MCP client to Parallel's native HTTP endpoint through mcp-remote.
• Preserve all existing providers, configuration, credentials, prompt hints, ordering, and defaults; the service runs only when a user selects it.

Validation

• One-file structural preservation check: passed.
• git diff --check: passed.
• Existing test suite was not run because npm is unavailable in the isolated worker environment.

Disclosure: I work at Parallel.ai, which operates this MCP endpoint.