dashscope

Pass

Audited by Gen Agent Trust Hub on Nov 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill integrates with Alibaba Cloud's DashScope API for multimodal tasks. It correctly recommends using .env files for secret management and mentions built-in redaction of API keys in logs, which is a best practice for API integrations.
  • [DATA_EXPOSURE_AND_EXFILTRATION]: The skill communicates with legitimate DashScope endpoints at dashscope.aliyuncs.com. These network operations are intended for service functionality and do not exhibit signs of exfiltrating sensitive local data.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data via prompts for image generation, text for speech synthesis, and external URLs for transcription (found in SKILL.md). While no explicit boundary markers are documented, the capabilities are restricted to service-specific API calls and writing to local output paths. The documented redaction logic in the tool's error handling provides mitigation against credential leakage.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 14, 2026, 08:42 AM