parallel-web

SUMMARY

What it does

This skill provides a unified interface to Parallel's web-intelligence workflows via the parallel-cli command-line tool. It routes user requests to one of several capabilities: Web Search for lookups and bounded research; Web Extract for fetching content from a known public URL (including PDFs and JavaScript-rendered pages); Data Enrichment for adding web-sourced fields to user-supplied rows; Deep Research for exhaustive, multi-source reports; FindAll for discovering entities matching natural-language criteria; and Monitor for recurring web change tracking. The skill includes setup instructions for installing parallel-cli (via uv tool install) and authenticating (interactively, via device flow, or using a PARALLEL_API_KEY environment variable). It emphasizes academic source priority for scientific queries, safety guidance for handling untrusted web content and constructing commands, context chaining via interaction IDs, and polling limits for long-running tasks. The skill is part of Scientific Agent Skills by K-Dense and includes citation instructions for the associated paper.