Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed workflow guide for spec-driven development and does not contain code, hidden execution, data access, or destructive behavior.
Before installing, confirm you want a Chinese-language workflow coordinator and review the separate superpowers and spec-kit dependencies it asks you to install from GitHub.
The argument hint allows broad inputs like 'describe the feature you want to build' or simply 'start', which creates an imprecise activation boundary. In an agent environment, vague invocations can cause the skill to trigger in unrelated contexts and steer the model into this workflow when the user did not clearly intend it, increasing the chance of unintended guidance or action sequencing.
The instruction requires the assistant to communicate in Chinese, which is a language/locale constraint. Because the file does not offer opt-in, user choice, or a clear region-specific justification, this is a natural-language policy violation under the stated rule.
The interaction rules say to proceed when the user says generic phrases like '开始' or '继续', but do not require confirmation that the user wants this specific skill. This can cause accidental takeover of conversations and misroute the agent into dependency-checking and workflow orchestration without a clear, task-specific request.
No suspicious patterns detected.
