← Back to Skills Marketplace
847
Downloads
0
Stars
2
Active Installs
2
Versions
Install in OpenClaw
/install perplexity-research
Description
Conduct deep research using Perplexity Agent API with web search, reasoning, and multi-model analysis. Use when the user needs current information, market re...
Usage Guidance
This skill appears to be a legitimate Perplexity API client for web research, but take these precautions before installing or enabling it: 1) Confirm the platform will request or store PERPLEXITY_API_KEY securely (manifest/SKILL.md require it; the registry header did not). 2) Review and consider removing or changing the hard-coded guidance in the client that tells the model to 'NEVER ask permission to search' — that encourages autonomous web searches which may query external sites without user consent. 3) Keep any .env files out of source control and backups; manifest excludes scripts/.env but double-check your packaging. 4) Inspect the included 'perplexity' dependency (pip package) if you want higher assurance about upstream behavior. If you are uncomfortable with the agent performing searches autonomously, do not enable the skill for autonomous invocation or ensure your agent's policy prevents unprompted web searches.
Capability Analysis
Type: OpenClaw Skill
Name: perplexity-research
Version: 2.0.0
The skill bundle provides a legitimate and well-documented interface for the Perplexity Research API. The core logic in `scripts/perplexity_client.py` is a standard API wrapper that handles queries, streaming, and model comparisons without any evidence of malicious behavior, data exfiltration, or unauthorized execution. While it references a non-existent model version (GPT-5.2), this appears to be a placeholder or developer error rather than a security risk. The requested 'network' permission in `manifest.yaml` is appropriate for the skill's stated purpose.
Capability Assessment
Purpose & Capability
The skill's code and SKILL.md implement a Perplexity API client with web search, streaming, multi-model comparison and cost tracking — all coherent with the name/description. However, the registry metadata at the top of the report lists 'Required env vars: none' while the SKILL.md and manifest explicitly require PERPLEXITY_API_KEY; this mismatch is an internal inconsistency that should be resolved.
Instruction Scope
The runtime instructions and the client code instruct the agent to perform web searches (web_search tool) and even include text telling the model to 'NEVER ask permission to search - just search when appropriate.' That guidance effectively encourages autonomous searching without explicit user confirmation and widens the skill's autonomy. Apart from that, the SKILL.md and code otherwise constrain activity to the Perplexity API and to reading a .env in the skill's scripts/ directory (expected for API keys).
Install Mechanism
There is no install spec (instruction-only skill), and included manifest/README recommend standard pip dependencies (perplexity, python-dotenv). No arbitrary downloads or extraction from untrusted URLs are present in the package contents provided.
Credentials
The skill asks for a single API credential (PERPLEXITY_API_KEY), which is proportionate to the declared functionality. However, registry metadata incorrectly lists 'none' for required env vars while manifest and SKILL.md require the key — this discrepancy can lead to surprises at install/runtime if the platform doesn't prompt for the secret. The code reads .env from the skill's scripts/ directory; ensure .env is not accidentally included in backups or shared.
Persistence & Privilege
The skill does not request 'always: true' and does not modify system-wide configurations. It only reads a local .env and otherwise uses the network to call the Perplexity API. Example workflows write reports to user-specified output paths (user-initiated).
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install perplexity-research - After installation, invoke the skill by name or use
/perplexity-research - Provide required inputs per the skill's parameter spec and get structured output
Version History
v2.0.0
Version 2.0.0
- Added explicit version number to SKILL.md metadata.
- No functional or usage changes in documentation content.
- All usage examples and integration patterns remain unchanged.
v1.0.0
Perplexity Research 1.0.0 introduces a research assistant leveraging the Perplexity Agent API:
- Enables deep research using web search, multi-model analysis, and advanced reasoning.
- Supports quick web searches, investment insights, trend analysis, and current information queries.
- Provides model comparison, streaming responses, and cost tracking features.
- Includes Python usage patterns, environment setup details, and command-line interface commands.
- Documents error handling, cost optimization tips, and best practices for integration.
Metadata
Frequently Asked Questions
What is Perplexity Research?
Conduct deep research using Perplexity Agent API with web search, reasoning, and multi-model analysis. Use when the user needs current information, market re... It is an AI Agent Skill for Claude Code / OpenClaw, with 847 downloads so far.
How do I install Perplexity Research?
Run "/install perplexity-research" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Perplexity Research free?
Yes, Perplexity Research is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Perplexity Research support?
Perplexity Research is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Perplexity Research?
It is built and maintained by Joe Hu (@hushenglang); the current version is v2.0.0.
More Skills