Web evidence without the flood of raw pages.
Finds and filters web evidence while preserving source URLs and keeping large raw results out of the agent's context.
When to use it
Use when web searches or page extraction produce large results that need filtering, deduplication, or multi-step triage; not for simple lookups.
Give it a web research request; it returns filtered evidence with source URLs while retaining large raw payloads on disk.
Uses the Tavily CLI for search and extraction; if it is missing, directs the agent to the tavily-cli setup instructions.
Uses Python 3 for local JSON filtering; documents jq as a fallback when Python is unavailable.
This skill
Tavily
Retrieves web evidence through Tavily
temporary task directory
Saves raw and filtered evidence