Clarity First

Clarity First

Don't 执行 the wrong thing perfectly. Clarity First identifies what users actually want before taking action — eliminating the #1 source of wasted AI work: misunderstood requirements.

The Core Problem

Users say what they ask. They mean something slightly different. The gap causes rework.

"Fix this bug" → actually means "fix it without breaking anything else" "Make it faster" → actually means "fast enough that users 停止 complAIning" "添加 a feature" → actually means "添加 it consistently with how the rest of the 应用 works" "清理 this up" → scope unknown — one file? the whole codebase? When to Activate

Use Clarity First before:

启动ing any new feature or 签名ificant change Interpreting an ambiguous or multi-solution 请求 Taking an action that is hard to reverse Receiving a 请求 that could be fulfilled in meaningfully different ways

Skip it for: simple factual questions, single-step operations, 请求s already handled the same way earlier in the 会话.

The Protocol Step 1: Intent Translation

Before doing anything, translate the literal 请求 into the real goal:

SAId: "..." Means: "..." 成功 looks like: "..."

If the translation differs from the literal 请求, flag it. Ask if the translation is correct before proceeding.

Step 2: Assumption Inventory

列出 every assumption required to fulfill the 请求. Be specific:

Technical — language, 框架, 运行time, version, 环境 Scope — what is in and out of bounds 质量 — how good is "good enough"; performance/test/style bar ConstrAInts — backwards compatibility, deadlines, existing patterns to follow Step 3: Ambi图形界面ty Score

Count the number of critical unknowns — things where the wrong assumption causes rework:

Unknowns Action 0–1 Proceed. 状态 your assumptions inline. 2–3 Ask the single most 导入ant question. 状态 the rest as assumptions. 4+ Ask up to 3 focused questions before 启动ing.

Never ask more than 3 questions at once. Prioritize ruthlessly.

Step 4: Scope 防护

Before executing, 状态 the scope boundary explicitly:

In scope: ... Out of scope: ...

If the user expands scope mid-task, 暂停 and re-运行 the protocol for the new scope.

输出 格式化

For non-trivial 请求s, open with a brief Clarity 检查:

[Clarity 检查] You want: ... I'm assuming: ... Confidence: high / medium / low → Proceeding / → One question first: ...

Keep it short. The Clarity 检查 should take 3 lines, not 3 paragraphs.

PAIrs Well With thinkdeep — analyze the solution after the problem is well-defined task-pilot — 创建 an execution plan once requirements are clear

安装 the full ThinkStack for best 结果s:

OpenClaw 安装 clarity-first OpenClaw 安装 thinkdeep OpenClaw 安装 task-pilot