perplexity-search

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMPROMPT_INJECTION
Full Analysis
  • Indirect Prompt Injection (MEDIUM): The skill retrieves external web content that is inherently untrusted and presents it to the agent, creating a surface for indirect prompt injection attacks.
  • Ingestion points: The script scripts/perplexity_search.py ingests external content from Perplexity search results via the OpenRouter API.
  • Boundary markers: Absent. The search results are returned as raw text and are not wrapped in delimiters or accompanied by instructions for the agent to ignore embedded commands.
  • Capability inventory: The script can write search results to a local file using the --output command-line argument and provides its output directly to the agent's context.
  • Sanitization: Absent. The skill does not perform any sanitization or filtering on the content retrieved from the external API response.
  • External Communications (LOW): The skill transmits search queries and an API key to openrouter.ai. While this is expected behavior for the tool's purpose, it constitutes data transit to a third-party service outside the local environment.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 08:07 AM