Perplexity Wrapped Search for Openclaw

A multi-mode AI search skill that integrates Perplexity's powerful web-grounded models directly into your agent workflows.

vacinc
v2.1.2
Feb 10, 2026
1
3k
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install perplexity-wrapped

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install perplexity-wrapped using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Perplexity Wrapped Search?

Perplexity Wrapped Search is a sophisticated tool designed for developers using Openclaw Skills to bridge the gap between LLMs and real-time web data. It provides three distinct operation modes: Sonar for conversational AI answers with citations, Search for traditional ranked results, and Agentic Research for leveraging third-party models like Claude or GPT with integrated web tools. By wrapping all outputs in security boundaries, it ensures that external content remains isolated from system instructions, providing a safe way to ingest live internet data into your AI environment.

Perplexity Wrapped Search Use Cases

  • Performing deep technical research with the sonar-deep-research model for comprehensive analysis.
  • Fact-checking claims using real-time web grounding and inline citations for verified accuracy.
  • Batching multiple search queries to gather diverse sources for data analysis and market research.
  • Integrating third-party LLMs like Claude or Gemini with live web access via the Agentic Research API.
  • Automating competitive analysis or news monitoring within an Openclaw Skills powered agent workflow.

How Perplexity Wrapped Search Works

  1. The user initiates a query via the CLI script, specifying the desired search mode, model, and optional parameters like reasoning level.
  2. The skill authenticates with the Perplexity API using a securely stored API key or 1Password integration.
  3. Depending on the mode, the skill either fetches ranked snippets, generates an AI response with citations, or orchestrates a multi-step research task using agentic tools.
  4. All retrieved data is processed through a security layer that sanitizes markers and normalizes characters to prevent content folding.
  5. The final output is returned inside untrusted-content boundaries to prevent prompt injection and maintain agent safety.

Perplexity Wrapped Search Setup

First, ensure you have Node.js installed. Configure your Perplexity API key in your configuration file to work with Openclaw Skills:

{
  "skills": {
    "entries": {
      "perplexity_wrapped": {
        "enabled": true,
        "apiKey": "pplx-your-key-here"
      }
    }
  }
}

You can then run a test query to verify the installation:

node scripts/search.mjs "latest developments in quantum computing"

Perplexity Wrapped Search Data Schema & Taxonomy

The skill organizes outputs based on the selected mode, ensuring clear attribution and structure for automated parsing.

Component Description
Security Boundary All output is wrapped in <<<EXTERNAL_UNTRUSTED_CONTENT>>> markers.
Main Content AI-generated response (Sonar/Agentic) or ranked snippets (Search).
Citations A structured list containing source titles and direct URLs.
Source Attribution Clearly identifies the origin as Web Search to distinguish from internal model knowledge.

Perplexity Wrapped Search Advanced Features

  • Support for Deep Research mode with comprehensive multi-step analysis for complex topics.
  • Agentic Research API integration allowing the use of specialized OpenAI, Anthropic, Google, and xAI models.
  • Customizable reasoning effort levels (low, medium, high) to balance cost and logic depth.
  • Batch query support in Search mode for high-efficiency data gathering in a single API call.
  • Integrated security features including boundary marker sanitization and content folding detection for safer Openclaw Skills operations.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Requires
Bins node
Github Stars: 0
forks: 0

Featured*