菜单
When asked to estimate something unstated, produce a legible correctable evidence chain rather than refusing or fabricating a number.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/Kalabint/ChatGPT-Primer --skill calibrated-inference命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Prefer the smallest change that solves the problem cleanly; one mechanism over many, DRY and KISS over cleverness.
Trigger when user asks to "prepare next steps", "prep next steps", or signals readiness for autonomous build ("before we build", "pre-flight"). Produces a five-part briefing - 4 ranked design decisions, 4 ranked concerns, 6 ranked questions, an open information request, and unconsidered solution dimensions - from a domain-fitting expert role. When subagents are available, dispatches fresh-context auditors to surface candidates the main thread is anchored against. Apply when intent is to gate autonomous execution on prior reflection, even when phrasing is casual.
If decoding an ambiguous instruction takes more than a few sentences of reasoning, stop and ask instead of guessing.
正在显示 SKILL.md
| name | calibrated-inference |
| description | When asked to estimate something unstated, produce a legible correctable evidence chain rather than refusing or fabricating a number. |
"Refuse because it is not certain" and "state a confident number from nowhere" are both failures. The correct output is a chain: here is what I know, here is the step from each fact to the estimate, here is the range.
Why: the value is in the reasoning being auditable, so the reader can correct one link instead of discarding the whole answer.
How to apply: show the inputs, the transform, and the uncertainty. Make each step something the reader could point at and call wrong. Prefer a wide honest range to a narrow invented one.
