ClawHub
Back to skill

Security audit

Perplexity Wrapped Search

Security checks across malware telemetry and agentic risk

Overview

This skill appears to do what it claims: it sends user search queries to Perplexity and returns results, with a raw JSON debug mode that needs care.

Install only if you are comfortable sending search queries, prompts, and optional agentic instructions to Perplexity, with agentic mode potentially involving third-party model providers. Use a dedicated API key, monitor costs, avoid submitting secrets or confidential data, and prefer the default wrapped output when another agent or automation will consume results.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (4)

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The CLI reference documents --json as 'raw JSON (debug mode, unwrapped)' but does not prominently warn that this bypasses the skill's main trust-boundary protection. In an agent or automation context, consumers may enable JSON for convenience and then inadvertently pass unwrapped model/web content into downstream components that treat it as trusted instructions or structured data.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The raw JSON example normalizes use of an unsafe mode without showing any warning about trust-boundary or privacy implications. Because this skill fetches external AI/web content, presenting --json as a simple debugging example can encourage reuse in production pipelines where unwrapped content may be ingested by agents, logged, or forwarded without adequate handling.

External Transmission

Medium
Category
Data Exfiltration
Content
}

  // Agentic Research API uses OpenAI Responses-style endpoint at /v2/responses
  const response = await fetch("https://api.perplexity.ai/v2/responses", {
    method: "POST",
    headers: {
      Authorization: `Bearer ${apiKey}`,
Confidence
79% confidence
Finding
fetch("https://api.perplexity.ai/v2/responses", { method: "POST"

External Transmission

Medium
Category
Data Exfiltration
Content
}

  // Agentic Research API uses OpenAI Responses-style endpoint at /v2/responses
  const response = await fetch("https://api.perplexity.ai/v2/responses", {
    method: "POST",
    headers: {
      Authorization: `Bearer ${apiKey}`,
Confidence
79% confidence
Finding
https://api.perplexity.ai/

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.