agentleFS
imMamdouhaboammarimMamdouhaboammar/kaku-chatgpt-harness/.agents/skills/perplexity/SKILL.md

perplexity

Skill0 starsChanged 4 months agoMITView on GitHub

Web search and research using Perplexity AI. Use when user says "search", "find", "look up", "ask", "research", or "what's the latest" for generic queries. NOT for library/framework docs (use Context7) or workspace questions.

9 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 6 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 6

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use ONLY when user says "search", "find", "look up", "ask", "research", or "what's the latest" for generic queries.

    line 8
  2. Distinctive:

    Use sparingly for complex questions only

    line 85
  3. Distinctive:

    Default to limited results: - avoid context bloat

    line 124
  4. Distinctive:

    Library docs = Context7: - ALWAYS try Context7 first

    line 125
  5. Distinctive:

    Deep research = /research: - NOT perplexity_research tool

    line 127
  6. Distinctive:

    Fallback chain: - Search → Ask → WebSearch (last resort)

    line 128

“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Perplexity Tools

Use ONLY when user says "search", "find", "look up", "ask", "research", or "what's the latest" for generic queries. NOT for library/framework docs (use Context7), gt CLI (use Graphite MCP), or workspace questions (use Nx MCP).

Quick Reference

Which Perplexity tool?

  • Need search results/URLs? → Perplexity Search
  • Need conversational answer? → Perplexity Ask
  • Need deep research? → Researcher agent (/research <topic>)

NOT Perplexity - use these instead:

  • Library/framework docs → Context7 MCP
  • Graphite gt CLI → Graphite MCP
  • THIS workspace → Nx MCP
  • Specific URL → URL Crawler

Perplexity Search

When to use:

  • Generic searches, finding resources
  • Current best practices, recent information
  • Tutorial/blog post discovery
  • User says "search for...", "find...", "look up..."

Default parameters (ALWAYS USE):

mcp__perplexity__perplexity_search({
  query: "your search query",
  max_results: 3,           // Default is 10 - too many!
  max_tokens_per_page: 512  // Reduce per-result content
})

When to increase limits: Only if:

  • User explicitly needs comprehensive results
  • Initial search found nothing useful
  • Complex topic needs multiple sources
// Increased limits (use sparingly)
mcp__perplexity__perplexity_search({
  query: "complex topic",
  max_results: 5,
  max_tokens_per_page: 1024
})

Perplexity Ask

When to use:

  • Need conversational explanation, not search results
  • Synthesize information from web
  • Explain concepts with current context

Usage:

mcp__perplexity__perplexity_ask({
  messages: [
    {
      role: "user",
      content: "Explain how postgres advisory locks work"
    }
  ]
})

NOT for:

  • Library documentation (use Context7)
  • Deep multi-source research (use researcher agent)

Prohibited Tool

NEVER use: mcp__perplexity__perplexity_research

Use instead: Researcher agent (/research <topic>)

  • Token cost: 30-50k tokens
  • Provides multi-source synthesis with citations
  • Use sparingly for complex questions only

Tool Selection Chain

Priority order:

  1. Context7 MCP - Library/framework docs
  2. Graphite MCP - Any gt CLI mention
  3. Nx MCP - THIS workspace questions
  4. Perplexity Search - Generic searches
  5. Perplexity Ask - Conversational answers
  6. Researcher agent - Deep multi-source research
  7. WebSearch - Last resort (after Perplexity exhausted)

Examples

✅ CORRECT - Use Perplexity Search:

  • "Find postgres migration best practices"
  • "Search for React testing tutorials"
  • "Look up latest trends in microservices"

✅ CORRECT - Use Perplexity Ask:

  • "Explain how postgres advisory locks work"
  • "What are the trade-offs of microservices?"

❌ WRONG - Use Context7 instead:

  • "Search for React hooks documentation" → Context7 MCP
  • "Find Next.js routing docs" → Context7 MCP
  • "Look up Temporal workflow API" → Context7 MCP

❌ WRONG - Use Graphite MCP instead:

  • "Search for gt stack commands" → Graphite MCP
  • "Find gt branch workflow" → Graphite MCP

❌ WRONG - Use Nx MCP instead:

  • "Search for build config" (in THIS workspace) → Nx MCP
  • "Find project dependencies" (in THIS workspace) → Nx MCP

Key Points

  • Default to limited results - avoid context bloat
  • Library docs = Context7 - ALWAYS try Context7 first
  • "gt" = Graphite MCP - ANY "gt" mention uses Graphite
  • Deep research = /research - NOT perplexity_research tool
  • Fallback chain - Search → Ask → WebSearch (last resort)

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.