perplexity-core-workflow-a

Pass

Audited by Gen Agent Trust Hub on Aug 21, 2026

Risk Level: SAFE
Full Analysis
  • [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill requires the openai package for both Node.js and Python environments to interact with the Perplexity API. These are well-known packages from an established service provider.
  • [DATA_EXPOSURE_AND_EXFILTRATION]: The skill accesses the PERPLEXITY_API_KEY from the environment. This follows best practices for secret management by avoiding hardcoded credentials. It communicates with the official Perplexity API endpoint at api.perplexity.ai.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content retrieved from web searches. Ingestion points: The answer and citations fields in the API response in SKILL.md. Boundary markers: None present in the code examples. Capability inventory: The examples are limited to formatting and logging data; they do not perform file writes or command execution based on the retrieved content. Sanitization: No explicit sanitization or validation of the search results is included in the templates.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 21, 2026, 09:39 PM
Security Audit — agent-trust-hub — perplexity-core-workflow-a