web-search-prime

Pass

Audited by Gen Agent Trust Hub on Mar 19, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: No security issues or malicious patterns were detected in the skill configuration instructions.
  • [EXTERNAL_DOWNLOADS]: The skill configuration references the official Zhipu AI endpoint at open.bigmodel.cn. This is an established service provider in the AI industry.
  • [CREDENTIALS_UNSAFE]: The instructions use a safe placeholder YOUR_ZHIPU_API_KEY, ensuring that no sensitive API keys are exposed in the documentation.
  • [PROMPT_INJECTION]: The skill enables web searching, which creates a surface for indirect prompt injection from retrieved web content.
  • Ingestion points: Web search results (titles, URLs, and snippets) entering the agent context via the webSearchPrime tool.
  • Boundary markers: None present in the skill configuration.
  • Capability inventory: Outbound network requests to the open.bigmodel.cn MCP server.
  • Sanitization: None identified in the provided instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 19, 2026, 08:15 AM