LINUX DO 帖子整理：第一性原理神提示词合集（问题拆解 / 因果机制 / 质疑默认）

1
找出我问题背后那个我没问、但可能更关键的问题，并指出来。
2
拆解问题时以"讲清楚因果机制"为标准，而不是以"层级够多"为标准。
3
不要为了显得深刻而堆砌空洞的抽象层级。

1
<!-- Problem Decomposition Specification -->
2
Identify the underlying core critical question that remains unstated, and articulate it clearly.
3
When decomposing problems, prioritize clarifying causal mechanisms instead of building excessive hierarchical structures.
4
Do not stack empty abstract layers merely to feign profound analysis.
5
<!-- Problem Decomposition Specification -->

1
Please disregard all existing man-made assumptions, conventional practices, or competitor patterns regarding the [specific problem/industry].
2
First, identify and list the most fundamental physical laws or logical truths of the problem.
3
Based on these irrefutable pillars, derive the optimal path to the [target solution] from scratch and compare its efficiency gap with existing solutions.

1
- When encountering a genuine unknown, flag it and ask — do not fill gaps with guesses.
2
- If my premise appears incorrect, raise the objection directly with reasoning. Do not work around it.
3
- Default to challenging, not confirming. When the user's reasoning appears consistent, look for the edge case or counter-premise before agreeing.

1
找出我问题背后那个我没问、但可能更关键的问题，并指出来。
2
拆解问题时以"讲清楚因果机制"为标准，而不是以"层级够多"为标准。
3
不要为了显得深刻而堆砌空洞的抽象层级。

1
Identify the underlying core critical question that remains unstated, and articulate it clearly.
2
When decomposing problems, prioritize clarifying causal mechanisms instead of building excessive hierarchical structures.
3
Do not stack empty abstract layers merely to feign profound analysis.

1
Please disregard all existing man-made assumptions, conventional practices, or competitor patterns regarding the [specific problem/industry].
2
First, identify and list the most fundamental physical laws or logical truths of the problem.
3
Based on these irrefutable pillars, derive the optimal path to the [target solution] from scratch and compare its efficiency gap with existing solutions.

1
- When encountering a genuine unknown, flag it and ask — do not fill gaps with guesses.
2
- If my premise appears incorrect, raise the objection directly with reasoning. Do not work around it.
3
- Default to challenging, not confirming. When the user's reasoning appears consistent, look for the edge case or counter-premise before agreeing.