invest-stock
Pass
Audited by Gen Agent Trust Hub on Nov 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions for the agent to execute a shell command:
python3 ~/.agents/skills/mx-data/scripts/get_data.py --type stock --code <代码> .... Since the<代码>placeholder represents user-supplied stock names or codes, there is a risk of command injection if the agent interpolates unsanitized input directly into the shell string without proper escaping. - [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves retrieving and processing untrusted external data, such as financial metrics, news, and corporate reports. This exposes the agent to indirect prompt injection, where instructions hidden in the data source could attempt to manipulate the agent's reasoning or behavior.
- Ingestion points: Data fetched via the
mx-datascript, including business descriptions and market news. - Boundary markers: The skill lacks explicit delimiters or instructions to ignore potential commands embedded within the retrieved financial data.
- Capability inventory: The skill allows shell command execution and local file access across its reference materials.
- Sanitization: There are no instructions for validating or sanitizing the external content before it is processed by the analysis framework.
Audit Metadata