kimi-websearch

Pass

Audited by Gen Agent Trust Hub on Jul 8, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill runs a local Python scrip t to manage search submission s and poll for result s from the vendor's API.
  • [EXTERNAL_DOWNLOADS]: The skill depen d s on the standar d request s librar y for all network communication.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it ingest s untrusted search data from external web site s.
  • Ingestion point s: Search result conten t provide d by script s/kimi_search.py.
  • Boundar y marker s: None; result s are displaye d withou t delimiter s or inst ruction s to ignor e embedde d command s.
  • Capabilit y inventor y: Local scrip t execution via python3.
  • Sanitization: The ret rieve d conten t is no t sanitize d or filtere d befor e bein g shown to the agen t.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 8, 2026, 11:21 AM
Security Audit — agent-trust-hub — kimi-websearch