spring-ai-alibaba
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [Prompt Injection] (LOW): The skill provides Java templates that demonstrate a surface for indirect prompt injection by directly interpolating user input into AI prompt templates. ● Ingestion points:
ChatService.chat(message)andPromptService.generatePrompt(style, question)inSKILL.md. ● Boundary markers: Absent in the provided code snippets. ● Capability inventory: The library (ChatClient) is designed to facilitate network communication with LLM APIs. ● Sanitization: No input validation or sanitization logic is included in the examples. - [Credentials Exposure] (SAFE): The documentation correctly guides users to use environment variables (
${DASHSCOPE_API_KEY}) for API keys, avoiding hardcoded secrets.
Audit Metadata