parallel-search

Use for current web search, URL extraction, or asynchronous deep-research task creation through Parallel.

SKILL.md

GitHub

Parallel Search

Exports parallel_search, parallel_extract, parallel_task, and parallel_task_result from src/index.ts.

All operations validate inputs before network access and return { ok: true, data } or { ok: false, error }. Task creation is nonblocking and returns the queued Task Run identifiers. parallel_task_result retrieves a run: wait: false polls its current status, wait: true blocks for the completed { status, output } and returns error.run_id on timeout so the run is never orphaned.

See Skill Operator Guide for Toolforge conventions.