| name | one-thing-focus |
| description | Identify the single highest-leverage action to focus on right now — from goals, projects, tasks, and constraints — so that other work becomes easier, unnecessary, or faster, and pair it with an explicit "not now" (based on The ONE Thing / bottleneck thinking). Use when the user has competing tasks or goals and needs to decide what to do next ("我现在最该做什么", "这几件事先做 哪个", "帮我聚焦", "我卡住了 / 没方向", "what's the one thing to focus on", "prioritize this", "find my first domino"). It is a focus / priority controller — it does NOT plan multi-step execution (a planner does that) and does NOT manage a todo list.
|
One Thing Focus
Purpose
Identify the single highest-leverage action that should receive the user's
primary attention right now.
The goal is not to optimize task completion.
The goal is to identify the action that, once completed, will make other actions
easier, unnecessary, or more effective.
Core Principle
Do not optimize for doing more. Optimize for doing the right next thing.
The test is not "is this important?" — it is:
After this is done, will other things become
- easier?
- unnecessary?
- faster to produce results?
- unblocked at the current bottleneck?
ONE Thing = Highest Leverage Action × Current Bottleneck × Strategic Alignment
Focus ≠ knowing what is important. Focus = actively ignoring what is not
important right now. Every ONE Thing must be paired with a NOT NOW.
Workflow
1. Define the desired outcome
Identify:
- What outcome is the user trying to achieve?
- What time horizon matters?
- What would success look like?
Do not select tasks yet.
2. Map current reality
Identify:
- Current state
- Desired state
- Major gaps
- Active projects
- Existing constraints
Separate symptoms, causes, and bottlenecks. Do not confuse a visible task with
the real gap.
3. Identify the current bottleneck
Ask:
What is currently preventing meaningful progress?
Prefer identifying a constraint or dependency instead of selecting the most
visible task. A good bottleneck has downstream consequences.
Four questions:
- 什么正在阻止目标前进？ — What is currently preventing meaningful progress?
- 哪个问题解决后影响最大？ — If solved, which problem would unlock the most progress?
- 哪些事情依赖它？ — What other tasks become easier if this is solved?
- 如果只能做一件事？ — If everything else had to wait, what should happen next?
4. Generate candidate actions
Generate 3–7 plausible actions. Each should directly address the bottleneck, a
critical dependency, or a strategic objective. Avoid generating a large task
list.
5. Evaluate leverage
Evaluate each candidate on:
- Impact — how much does completion matter?
- Unlocking power — how much downstream work does it unlock?
- Bottleneck fit — does it directly relieve the current bottleneck?
- Strategic alignment — does it advance the core goal?
- Effort — is the cost reasonable?
- Time sensitivity — is now the right moment?
Prefer actions that create downstream progress.
6. Select ONE Thing
Select exactly one action. Do not return multiple equal priorities.
ONE THING:
[Single highest-leverage action]
WHY:
[Why this is the current first domino]
NOT NOW:
[Important tasks intentionally deferred]
Never output "你可以同时做 A、B、C." Output "现在最应该做的是 A."
7. Find the First Domino
Break the ONE Thing into a concrete next step.
The First Domino must:
- be executable immediately
- take less than 30 minutes to start
- have a clear completion condition
- require no further prioritization
Avoid vague actions: research, think about, improve, work on.
Prefer concrete actions: create, define, implement, compare, write, test.
8. Protect focus
Recommend a time block. During the block: defer unrelated tasks, capture
distractions, avoid reprioritizing, and focus only on the First Domino.
9. Review
After execution, ask:
- What changed?
- Was the bottleneck reduced?
- What was discovered?
- Has the bottleneck moved?
- What is the next ONE Thing?
Do not automatically reuse the previous priority. Recalculate from the new
reality.
The Execution Loop
ONE Thing is not permanent. Finishing one bottleneck changes the system state,
so the loop restarts.
FOCUS — what is the ONE Thing?
↓
COMMIT — block the time
↓
EXECUTE — do the First Domino
↓
REVIEW — did reality move?
↓
REFRESH — find the next ONE Thing
Day 1: bottleneck = 没有 Execution Trace → ONE Thing = 建立 Trace Schema
Day 5: bottleneck = 有 Trace 但无法分析失败 → ONE Thing = 建立 Failure Classification
This is a dynamic priority loop, not a static priority list.
Output Format
━━━━━━━━━━━━━━━━━━
🎯 DESIRED OUTCOME
[目标，一句话]
━━━━━━━━━━━━━━━━━━
🔍 CURRENT BOTTLENECK
[当前真正卡住的地方]
━━━━━━━━━━━━━━━━━━
🥇 ONE THING
[唯一最高杠杆动作]
━━━━━━━━━━━━━━━━━━
WHY THIS
它将直接解锁：
→ ...
→ ...
━━━━━━━━━━━━━━━━━━
🚫 NOT NOW
- ...
- ...
（这些事情并不消失，只是暂时不是当前的 First Domino。）
━━━━━━━━━━━━━━━━━━
▶ FIRST DOMINO
[现在即可执行的第一个具体动作]
完成条件：
[ ] ...
[ ] ...
━━━━━━━━━━━━━━━━━━
⏱ FOCUS BLOCK
[时长]
Only objective:
[本次唯一目标]
━━━━━━━━━━━━━━━━━━
