perplexity-architecture-variants

Pass

Audited by Gen Agent Trust Hub on Aug 26, 2026

Risk Level: SAFE
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill correctly demonstrates using process.env.PERPLEXITY_API_KEY and os.environ["PERPLEXITY_API_KEY"] for secret management, avoiding hardcoded credentials.
  • [EXTERNAL_DOWNLOADS]: The skill references official documentation and pricing from Perplexity's official domain (docs.perplexity.ai), which is a well-known service.
  • [DATA_EXFILTRATION]: No sensitive file access or unauthorized network operations were detected. Network calls are directed to the official Perplexity API endpoint.
  • [COMMAND_EXECUTION]: The skill does not contain any shell command execution patterns or privilege escalation attempts.
  • [INDIRECT_PROMPT_INJECTION]: While the skill handles user-provided queries, it is a code blueprint provider rather than an active data processing agent, and the provided code examples use standard API request structures.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 26, 2026, 07:54 AM
Security Audit — agent-trust-hub — perplexity-architecture-variants