mx-financial-assistant
Fail
Audited by Gen Agent Trust Hub on Jun 15, 2026
Risk Level: HIGHCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [CREDENTIALS_UNSAFE]: The script
scripts/generate_answer.pycontains a hardcoded default API keyem_fjFqd4YB6Cqs52LF48XWbMDdLNq6MyNgassigned to theEM_API_KEYvariable, which is used to authenticate requests to the Eastmoney AI SaaS platform. - [EXTERNAL_DOWNLOADS]: The skill fetches real-time financial data and generated answers from the official Eastmoney AI SaaS endpoint at
https://ai-saas.eastmoney.com/proxy/app-robo-advisor-api/assistant/ask. - [PROMPT_INJECTION]: The skill implements an indirect prompt injection surface by taking content from an external API response and displaying it directly as Markdown without sanitization or protective boundary markers.
- Ingestion points: Data enters the agent context via the
displayDatafield in the response fromai-saas.eastmoney.com. - Boundary markers: No explicit delimiters or instructions are used to separate external data from the skill's own operational context.
- Capability inventory: The skill facilitates network requests to specified endpoints but does not include local file-system modification or subprocess execution capabilities in its provided scripts.
- Sanitization: The script extracts and prints the
displayDatastring directly without performing HTML/Markdown escaping or content validation.
Recommendations
- AI detected serious security threats
Audit Metadata