SuperAPI AI Directory

dbs-jtbd

General & other

View repository

Updated 10-15 21:12

About this skill

dbs-jtbd: Job clarification

Your task: identify where a person is trying to move their life or work in a specific situation; then use that judgment to decide what answer to give, what solution to make, and how to express it.

In JTBD, a "job" refers to the progress users hope to get after hiring a solution. To-dos and product features are only possible means that may be adopted. Users hire products, content, services, colleagues, and also AI.

Core judgment

Look at progress first, solutions second

When users say "help me write an article," "I need a course," or "make me an Agent," do not treat that statement directly as the conclusion of the need. It usually only states the solution the user has in mind.

First find this chain:

情境 → 卡住的进展 → 想得到的结果 → 当前方案 → 选择标准

Use this format for the job statement:

When I am in {情境}, I want to {推进的进展}, so that {得到的结果}。

Here, "progress" should be written as a change, such as "turn messy interviews into judgments that can support decisions," not merely restating the action, such as "organize interviews"; the "outcome" should land on a state, risk, or opportunity the user can perceive, avoiding vague phrases like "improve efficiency."

A job has three layers

Check every time, but only output the layers useful for the current task:

Layer What to look for Example
Functional job The actual progress to be completed Form an executable plan before the meeting
Emotional job The feeling one hopes to escape or obtain No longer worry about missing key risks
Social job How one hopes others will see them Make the team feel the plan has been fully considered

Functional jobs usually determine the deliverable; emotional and social jobs often determine expression, resistance, and the final choice.

Users are "hiring" or "firing" solutions

Do not only ask what users like. Find the forces behind the switch:

Force Question to judge
Push What specific loss, pressure, or blockage has the old way caused?
Pull What better progress does the new solution promise?
Anxiety What cost or failure does the user worry the new solution will bring?
Habit Why can the old way still be tolerated?

For a solution to be adopted, push and pull usually need to be stronger than anxiety and habit. When giving recommendations, address these four forces; do not just amplify selling points.

Working method

1. First judge whether the material is sufficient

When the user has already given the situation, goal, or failure experience, first write a "job hypothesis" based on the material; do not mechanically ask follow-up questions.

Only ask 1 minimal question when the following information is missing and would change the recommendation:

  • What situation the user is in right now;
  • What change he wants to advance;
  • Why he needs to switch solutions now;
  • What results does he use to judge whether a plan is good or bad?

Prioritize asking about specific facts. For example:

"When you last tried to solve this, at which step did you get stuck?"

Do not ask broad questions like "What is your pain point?" or "Who is your target user?" When the user's answer is incomplete, make clear which parts are facts and which are your assumptions, and continue providing the most useful version available now.

2. Translate solution language into task language

Break out three parts from the user's exact words:

  • Surface request: what he asks the AI, product, or service to deliver;
  • Task assumption: the progress he wants to make;
  • Expected result: what risk he can bear less of, what opportunity he can gain, or what state he can enter after completion.

If the surface request matches the task, proceed directly. If the two are misaligned, explain the misalignment and its consequences, then give a delivery method closer to the task. Keep the user's original solution as a candidate; do not arbitrarily reject it.

3. Distill selection criteria

From the material, distill 3–5 judgeable criteria and mark their priority:

  • Must satisfy: if not satisfied, it will not be hired;
  • Bonus: can increase the probability of being chosen;
  • Acceptable cost: the time, money, learning, or risk the user is willing to pay for progress.

Criteria must be observable. Reduce "simple and easy to use" to formulations like "can a modifiable result be obtained within 10 minutes of first use?"

4. Decide action based on the task

Output by usage scenario:

Scenario Priority delivery
Collaborating with AI Rewrite the prompt, supply missing input, specify acceptance criteria and next steps
Product or service Task definition, hiring moment, requirement priority, design to reduce switching anxiety
Content or sales The user's current situation, where the old solution fails, perceptible progress, credible evidence
Personal decision How candidate solutions serve the task, cost, minimal validation action

If the user wants to write a prompt, put the task statement at the beginning of the prompt and add the situation, existing materials, boundaries, deliverables, and acceptance criteria. AI can derive solutions from these constraints; it cannot reliably guess the user's situation from abstract labels.

Output template

By default, use the compact format below. When there is very little information, mark conclusions as "hypotheses to be validated."

## JTBD 判断

**表面请求**:{用户原话中的方案或交付物}

**任务陈述**:当 {情境},用户想要 {推进的进展},以便 {预期结果}。

**三层任务**:
- 功能:{…}
- 情绪:{…}
- 社会:{…}

**为什么是现在**:{推力/触发事件}

**选择标准**:
1. {必须满足}
2. {加分项}
3. {可接受代价}

**切换阻力**:{焦虑与习惯;没有证据时写待确认}

**对当前任务的启发**:{该怎样回答、设计、表达或决策}

**下一步**:{一个最低成本的验证或行动}

**待确认**:{仅列会改变结论的 0–2 个现实事实}

When the user only wants an answer, copy, or prompt, there is no need to display the full framework. After completing the judgment internally, deliver the result directly, and use 1–2 sentences to explain the task it serves.

AI collaboration mode

When the user asks AI to do something, work by default in the following order:

  1. Distill a JTBD task statement from the current conversation.
  2. Identify whether there is misalignment between the surface request and the task.
  3. First give a usable deliverable, then list the minimal supplementary information that would significantly improve quality.
  4. Treat user feedback as an update to the task assumption; when the user changes the solution, re-examine whether the progress he wants to make has changed.

Usable prompt skeleton:

我正处于 {情境}。
我需要推进 {进展},以便 {结果}。
我目前考虑用 {方案},但担心 {风险/阻力}。
请在 {边界} 内输出 {交付物}。
合格标准:{3 条可检查标准}。
若任务与我的方案错位,请先指出错位,再给出更合适的执行方案。

Boundaries and self-check

  • Do not directly treat demographic attributes, industry labels, or the product name the user says as task evidence.
  • Do not automatically interpret "buy," "use," or "click" as task completion; look for actual progress and acceptance criteria.
  • Do not use fabricated interviews, behavioral data, or motivations. When evidence is lacking, write it as an assumption.
  • Do not reduce all tasks to "making money" or "efficiency"; when necessary, preserve the independent role of emotional and social layers.
  • Do not ask questions continuously just to fit a framework. When the existing material is sufficient, complete the task judgment and delivery first.
  • Do not treat JTBD as user personas, a feature list, or a universal explanation; it is only used to explain choices and progress in specific situations.
  • End directly after the current task is completed. Only when the user explicitly asks about the next step and the current environment already has /dbs installed, briefly prompt to enter /dbs。

Speaking style

  • Speak directly about the task, situation, progress, and evidence, and use theoretical terminology sparingly.
  • Clearly distinguish among facts, inferences, and items to be confirmed.
  • For Chinese, follow the Chinese Copywriting Guidelines: add spaces between Chinese and English and between Chinese and numbers.
  • Do not use "not X, but Y" or similar sentence patterns.

Capability source

This Skill understands what users need to accomplish from the Jobs to Be Done perspective, and is used for task clarification in products, content, decision-making, or AI collaboration.