The financial-ai-analyst skill is vulnerable to
Claims to do
金融资讯搜索: 通过**自然语言查询**检索时效性金融信息,数据来自于妙想大模型,适用场景包括: - **最新新闻与政策动态** - **公司公告与事件跟踪** - **券商研报与市场解读** - **宏观事件对市场/板块影响分析**
Actually does
The skill executes a Python script (`scripts/get_data.py`) that uses an `EM_API_KEY` to query an API provided by East Money's Miaoxiang service (likely via `https://ai.eastmoney.com/mxClaw`). It retrieves financial information like news, announcements, and research reports. The retrieved content is then printed to the console and can be optionally saved as a `.txt` file in the local workspace.
openclaw skills install financial-ai-analyst/mx-finance-searchDeBERTa classifier detected prompt injection (confidence: 1.00)
The skill executes a Python script with user-provided input as a command-line argument. If the underlying script does not properly sanitize this input before using it in shell commands, it could lead to command injection.
python3 {baseDir}/scripts/get_data.py "寒武纪 688256 最新研报与公告"The skill supports saving query results to local .txt files, with an `output_dir` parameter. If the directory path or filename generation is vulnerable to path traversal, an attacker could write files to arbitrary locations on the system.
支持将结果保存为本地 `.txt` 文件, output_dir=Path("workspace/mx_finance_search")[](https://mondoo.com/ai-agent-security/skills/clawhub/financial-ai-analyst/mx-finance-search)<a href="https://mondoo.com/ai-agent-security/skills/clawhub/financial-ai-analyst/mx-finance-search"><img src="https://mondoo.com/ai-agent-security/api/badge/clawhub/financial-ai-analyst/mx-finance-search.svg" alt="Mondoo Skill Check" /></a>https://mondoo.com/ai-agent-security/api/badge/clawhub/financial-ai-analyst/mx-finance-search.svgSkills can read files, run commands, and access credentials. Mondoo helps organizations manage the security risks of AI agent skills across their entire fleet.