Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed remote workflow for generating reducer CAD outputs, with notable privacy considerations but no evidence of hidden or malicious behavior.
Install only if you are comfortable sending reducer design parameters to jixietools.com and receiving guest links that may grant access to status or output files. Treat filename, guest_code, and generated links as sensitive workflow tokens, and confirm before creating a production sheet or sharing links.
Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.
curl -s "https://jixietools.com/api/v1/products?category_id=8" | python3 -m json.tool
Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.
curl -s "https://jixietools.com/api/v1/products?category_id=8" | python3 -m json.tool
Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.
用户选择后，获取该产品的参数定义：
curl -s "https://jixietools.com/api/v1/products/PRODUCT_ID/start" | python3 -m json.tool
返回结构：
Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.
curl -s -X POST "https://jixietools.com/api/v1/products/PRODUCT_ID/calculate" \
-H "Content-Type: application/json" \
-d '{"inputs": {"修改的参数名": "新值"}, "filename": "之前保存的filename"}'
清单与说明多次宣称该技能用于“减速器CAD图纸生成”以及“完成减速器图纸的生成流程”，但具体步骤是调用产品参数计算、创建 production sheet、再轮询制作状态与输出文件。文档中的实现意图更接近“提交远程制作任务并查看结果”，与“直接生成CAD图纸”的表述存在语义偏差。
The description says to invoke the skill when the user says phrases including “减速器” and “做一个减速器”. These are broad natural-language expressions that could appear in casual discussion or brainstorming, but the file does not provide constraints or negative examples to distinguish mention from intentional invocation.
This step sends user-supplied or workflow-derived data to an external service to precompute dropdown options, but the skill does not require informed consent or data-minimization before transmission. Even if the payload is small, external transmission of design parameters to a third party can expose proprietary engineering inputs.
curl -s -X POST "https://jixietools.com/api/v1/products/PRODUCT_ID/calculate" \
-H "Content-Type: application/json" \
-d '{"inputs": {"有选项的参数名": ""}}'
The incremental update call transmits changed parameters plus the persistent filename token to the external API, tying a user's session state to a remote workbook. If mishandled, the filename can act as a capability identifier for modifying or correlating design state across requests.
curl -s -X POST "https://jixietools.com/api/v1/products/PRODUCT_ID/calculate" \
-H "Content-Type: application/json" \
-d '{"inputs": {"修改的参数名": "新值"}, "filename": "之前保存的filename"}'
The skill expands from CAD assistance into unauthenticated order/workflow creation and purchase funneling, which exceeds the declared scope and can trigger external side effects without strong user awareness. This increases the risk of unintended submissions, privacy surprises, and abuse of the guest workflow endpoints.
The skill exposes guest-access production links and instructs polling of unauthenticated status endpoints without warning the user that possession of the link/code may grant access to order status and outputs. If these guest URLs or codes are leaked through logs, chat history, or shoulder surfing, third parties may monitor or retrieve sensitive design artifacts.
The interaction rules explicitly require the agent to communicate in Chinese for the entire workflow. This is a language-policy constraint applied unconditionally, with no user choice or opt-in documented.