financial-expert
Pass
Audited by Gen Agent Trust Hub on Aug 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill relies on the
bl(Bailian CLI) to execute tools via the Model Studio MCP marketplace (market-cmapi00073529). It uses shell commands to fetch structured and non-structured financial data. - [EXTERNAL_DOWNLOADS]: The documentation instructs the user to install the
bailian-clipackage from the NPM registry and references official Alibaba Cloud documentation for installation and authentication. - [CREDENTIALS_UNSAFE]: The skill provides instructions for managing the
DASHSCOPE_API_KEY. It recommends secure methods such as using environment variables (export DASHSCOPE_API_KEY) or the CLI's own authentication command (bl auth login). - [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources, including analyst research reports, company announcements, and financial metrics. This data is interpolated into the agent's context for summarization and analysis, representing an indirect injection surface.
- Ingestion points: Data returned from MCP tools like
FinancialResearchReportandAnnouncementData, as well as user-provided queries. - Boundary markers: The skill defines specific output presentation rules (tables, summaries) and advises against inventing data, but lacks explicit data sanitization or boundary markers between external content and instructions.
- Capability inventory: The agent has the capability to execute shell commands through the
blCLI. - Sanitization: No explicit sanitization or filtering of external data is mentioned in the skill instructions.
Audit Metadata