web-search-perplexity

Pass

Audited by Gen Agent Trust Hub on Jun 23, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses python3 to parse and extract the API key from a local configuration file at ~/.gooseworks/credentials.json.
  • [EXTERNAL_DOWNLOADS]: Setup instructions specify using npx gooseworks login, which downloads the vendor's utility from the NPM registry to initialize credentials.
  • [DATA_EXFILTRATION]: The skill transmits search queries and model parameters to api.gooseworks.ai, which is the official API proxy domain for the skill's author.
  • [PROMPT_INJECTION]: The skill retrieves real-time web search results from Perplexity, introducing an attack surface for indirect prompt injection.
  • Ingestion points: Untrusted web content retrieved via the Perplexity API is processed within the agent's context.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the provided templates.
  • Capability inventory: The skill is capable of performing network requests via curl to the vendor's proxy.
  • Sanitization: No sanitization or content filtering logic is described for the retrieved search results.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 23, 2026, 02:15 PM
Security Audit — agent-trust-hub — web-search-perplexity