plugin_manifest: id: webpilot-crawler-v2 name: "WebPilot: The Hydra Crawler" version: "2.1.0" tagline: > An AI-native web intelligence engine. Maximizes token efficiency, automates deep-dives, and provides structured dataβnot just raw HTML. Unleash the Hydra.
config_notes: MAX_CONTENT_SIZE: "50000 characters" CHUNK_SIZE: "4000 characters (Optimized for context windows)" AUTO_CRAWL_LIMIT: "Top 3 search results"
features:
title: Auto-Pilot Search π detail: The primary searchWeb tool executes a simultaneous, shallow crawl of the top 3 results, turning one query into multiple deep insights. icon: ποΈ
title: Context Scroller π detail: Navigate huge documents effortlessly. The scrollPage tool delivers content in digestible, token-friendly 4K chunks, conserving your LLM's memory. icon: π§
title: Link De-Obfuscation π
detail: All embedded links are surgically converted to an explicit, actionable format: [Link Text (https://full.url)], giving the AI crystal clarity.
icon: ποΈ
title: Token Efficiency detail: Aggressive payload slimming ensures near-zero filler, delivering maximum useful information per API call. icon: β‘
tools:
name: searchWeb alias: One-Shot Research description: Executes the core search and auto-crawl sequence. parameters:
name: crawlWebPage alias: Session Initialization description: Manually starts a session for a deep dive on a specific URL, activating the scroller state. parameters:
name: scrollPage alias: Navigator Protocol description: Fetches the next sequential chunk of content from an active crawl session. parameters: