zhihu-search

Pass

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted data from external search results via the zhihu_search, global_search, and zhida tools. This presents a surface for indirect prompt injection where malicious instructions embedded in web content could influence the agent's behavior.
  • Ingestion points: Results from Zhihu and global web searches (SKILL.md).
  • Boundary markers: No delimiters or specific instructions to ignore embedded prompts are provided in the skill logic.
  • Capability inventory: Network access via search tools and file-write access for credential persistence (set_zhihu_api_key).
  • Sanitization: The instructions do not specify any validation or sanitization of the retrieved content.
  • [SAFE]: The skill requires a user-provided Access Secret for API authentication. It uses the set_zhihu_api_key tool to persist this credential in a local data directory. This follows standard secret management practices for integrations and does not involve hardcoded secrets or unauthorized exfiltration.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 18, 2026, 10:06 AM
Security Audit — agent-trust-hub — zhihu-search