By allowed-tools:
- Task
- Read
- Grep
- Glob
- WebFetch
- mcp__perplexity-ask__perplexity_ask
argument-hint:
context: fork description: Web research via platform web tools. Use for technical comparisons, recent facts, ecosystem news, best practices, standards, or questions needing grounded web evidence. NOT for API syntax lookup or code examples — use looking-up-docs for those. NOT for repo-specific questions — search local files first. name: researching-web user-invocable: true
Web Research
Follow the base skill. This Claude overlay only defines tool use and execution details.
Claude tool rules
- Use Perplexity for broad research questions, comparisons, or source discovery when it is available.
- Use
WebFetchto inspect primary or official sources directly after source discovery. - Do not answer research questions from memory.
- Do not route exact API syntax or code-example questions here; use
looking-up-docs. - Do not route repo-specific questions here when local files can answer them.
Tool selection on Claude
- Simple factual question: use the smallest web query that can answer it.
- Source selection: use Perplexity or another available search-capable tool, then fetch the best sources.
- Broad investigation: use Perplexity first when it fits; if the runtime offers a deeper research mode, use it when the question truly needs it.
- Follow-up detail: fetch only the sources needed to support the claim.
Inspect enough sources to support the answer. Do not fetch every citation unless the risk or ambiguity justifies it.
Source handling
- Prefer official docs, standards, vendor docs, maintainer posts, and primary announcements.
- Treat forum summaries and secondary blogs as supporting evidence, not sole authority, when stronger sources should exist.
- If sources conflict, show the conflict and reduce confidence.
- If only a research plan is possible, say that final evidence must be grounded in URL-cited sources.
Optional delegation
Use Task only for broad background research that can run independently without
changing the output contract. Keep the prompt bounded and ask it to return cited
facts, not final recommendations.
Output
Use the base skill's Research Result output contract exactly.