Lp3
Medium
- Category
- MCP Least Privilege
- Confidence
- 88% confidence
- Finding
- The skill instructs the agent to invoke a local shell command (`qg`) and generate external WeChat links, which are code-execution and network-adjacent capabilities without any declared permissions or trust boundary documentation. Even though the described workflow is limited to schedule lookup and link generation, undeclared shell/network capability increases the risk of command misuse, unexpected side effects in the local CLI, or data exposure if the underlying tool behaves differently than assumed.
