perplexity — Claude Code skill: web search and research
This skill wires Perplexity into Claude: web search, conversational answers, and deep research through an agent. Handy when you need current information rather than library docs.
npx claude-code-templates@latest --skill ai-research/perplexityWhat perplexity does
The skill tells the agent which tool to reach for. Plain lookups go to mcp__perplexity__perplexity_search, explanations grounded in the web go to mcp__perplexity__perplexity_ask, and deep dives go to the /research <topic> agent. The mcp__perplexity__perplexity_research tool is off-limits here — the skill points you at the researcher agent instead.
Searches are meant to run tight: max_results: 3 and max_tokens_per_page: 512, since the defaults bloat context. Raise those only when the first pass came back empty or the topic genuinely needs several sources.
Scope is the real constraint. Library and framework docs belong to Context7 MCP, anything mentioning gt belongs to Graphite MCP, and questions about the current workspace belong to Nx MCP. Mix those up and your query hits the open web instead of exact docs — you will notice it in the answers. Running this requires the Perplexity MCP server with an API key configured.
What it helps with
- find current best practices that local docs do not cover
- get a conversational explanation backed by web sources
- run a multi-source deep dive on a topic
- route a query to the web instead of framework documentation
- keep context small during web searches
Example prompts for the agent
- find recent articles on postgres migrations
- explain how postgres advisory locks work
- run /research on microservices trade-offs
How to install perplexity
1claude-code-templates
npx claude-code-templates@latest --skill ai-research/perplexity2Skills CLI (Vercel)
npx skills add davila7/claude-code-templates --skill perplexity -a claude-code3Manually: put the files in place
~/.claude/skills/perplexity/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/perplexity/
Manually: put the files in place
.agents/skills/perplexity/SKILL.mdFor all projects use ~/.agents/skills/
Manually: put the files in place
.cursor/skills/perplexity/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install davila7/claude-code-templates perplexity2Manually: put the files in place
.github/skills/perplexity/SKILL.mdPlatform CLI
gemini skills install https://github.com/davila7/claude-code-templates --path cli-tool/components/skills/ai-research/perplexityManually: put the files in place
.devin/skills/perplexity/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about perplexity
How do I install perplexity in Claude?
Use the templates CLI: npx claude-code-templates@latest --skill ai-research/perplexity. Alternatively run npx skills add davila7/claude-code-templates --skill perplexity -a claude-code. The skill file lands at ~/.claude/skills/perplexity/SKILL.md.
What does the perplexity skill do?
It spells out for the agent when to call each Perplexity tool: search, conversational answer, or deep research. It also separates web queries from documentation queries so the agent stops searching the web for framework docs.
How is perplexity different from Context7?
The skill puts Context7 first for library and framework documentation. Perplexity comes next and takes general queries: trends, tutorials, recent news. Workspace questions go to Nx MCP.