pm-okr

Preamble (run first)

if [ -f "docs/05-产品战略/战略决策建议.md" ]; then
  echo "✅ 战略决策建议 - 已找到（可承接为 O 的来源）"
else
  echo "⏳ 战略决策建议 - 未找到（可选）"
fi

---

Write-time boundaries

When only discussing, do not write files, and the startup check does not create directories. When the user asks to save artifacts, first confirm the current product, the exact output path, and the overwrite/append choice for existing files; after confirmation,only immediately before the actual Writecreate the parent directory of that output file. For full installation, see skills/_shared/artifact-policy.md, and installing this skill alone also follows this paragraph.

---

Cross-Agent interaction rules

When the process requires interaction with the user:

1. If the current environment supports AskUserQuestion, use AskUserQuestion (best experience).
2. If the current environment does not support AskUserQuestion, you must ask the same question via a normal chat message.
3. Ask only one question at a time.
4. After asking, you must stop the current turn and wait for the user's answer (STOP and WAIT).
5. Do not generate documents or write to docs before the user answers.
6. Existing docs files cannot substitute for the user's answer in this turn.

---

Applicable scenarios

• The user says "set OKRs" "goal breakdown" "how to write key results" "goal alignment" "quarterly goals"
• Distinguish from pm-iteration(iteration to-do): OKRs managedirection and measurable results, iteration managesexecution scheduling。

---

OKR fundamentals (brief)

• O（Objective）: directional, inspiring, qualitative, no piling up of metrics.
• KR（Key Result）: measurable, with baseline and target values, verifiable, usually 2-4 per O.
• Alignment (Cascading): company → product/team → individual, lower-level KRs support upper-level Os.
• Confidence index: mark 50% (challenging) / 70% (fairly stable) at formulation time, used for review.

---

Execution flow

Step 1: Clarify the source of goals (main agent - user interaction)

Use AskUserQuestion to ask:

🎯 OKR scope

What level does this OKR land at?

A) Product-level OKR (inheriting company/strategy) B) Team-level OKR C) Individual/project-level OKR

Goal source:

1. Follow the strategic decision recommendations
2. I will manually input business priorities
3. Derive based on the current growth/data status

Record OKR_LEVEL and OBJECTIVE_SOURCE。

---

Step 2: Distill the Objective (main agent)

Based on the source, distill 1-3 Os, requiring:

• One-sentence direction, inspiring, containing no metrics
• Consistent with the strategy/business priorities

Example: "Make the product the first choice among target users" "Get the growth engine running and sustainable"

Use AskUserQuestion to confirm the wording and number of Os with the user.

---

Step 3: Break down into Key Results (main agent + optional subagent)

Break down 2-4 KRs for each O, following:Measurable + baseline → target + verification method。

Optionally use the Agent tool to dispatch a subagent to generate KR drafts for complex topics:

Tool: Task
Parameters:
  subagent_type: "general-purpose"
  description: "OKR-KR草案"
  prompt: |
    你是目标管理专家。为以下 Objective 拆解 Key Results。
    O: {O}
    上下文: {产品现状/数据}
    要求：每个 O 2-4 个 KR；每个 KR 含 基线值→目标值、度量方式、信心指数(50%/70%)；避免把任务当 KR。
    输出 Markdown。

The main agent integrates and confirms with the user.

---

Step 4: Generate the alignment map and review cadence (main agent)

• Alignment Map: Support relationship between upper-level O and lower-level KR
• Review Cadence: Biweekly check-in progress, quarterly review scoring (0-1.0, 0.7 is the passing threshold)
• Risks and Dependencies: Dependencies between KRs, resource requirements

---

Step 5: Generate OKR document (main agent)

Use the Write tool to generate docs/03-增长迭代/OKR目标管理.md：

# {产品/团队} OKR（{周期}）

## 一、Objective 与 Key Results
### O1: {方向性目标}
- KR1: {基线} → {目标}，度量: {方式}，信心: {50%/70%}
- KR2: ...

### O2: {方向性目标}
- KR1: ...

## 二、对齐地图
| 上层 O | 支撑的下层 KR |
|-------|--------------|
| {公司O} | {产品KR1, KR2} |

## 三、度量与基线
| KR | 当前基线 | 目标 | 数据源 | 负责人 |
|----|---------|------|-------|-------|
| {KR} | {值} | {值} | {源} | {人} |

## 四、复盘节奏
- 双周: 进度打卡
- 季度: 评分复盘（0-1.0，0.7 达标）

## 五、依赖与风险
- 依赖: {列表}
- 风险: {列表}

## 六、下一步建议
1. /pm-iteration - 将 KR 转为迭代待办
2. /pm-retro - 季度复盘
3. /pm-report - 跟踪 KR 进度

---

Step 6: Recommend next steps (main agent)

✅ OKR generated:docs/03-增长迭代/OKR目标管理.md

Suggested actions:

1. /pm-iteration - Convert KRs to iteration to-dos
2. /pm-retro - Quarterly review
3. /pm-report - Progress tracking

---

Output Quality Comparison

✅ Good Example：

• Measurable KR: "MAU 120k → 200k, data source GA, confidence 50%"
• O contains no metrics: "Significantly improve new user first-week retention"
• Has alignment and review

❌ Bad Example：

• Treating a task as a KR: "Complete the launch of the new version"
• O written as a pile of metrics: "Increase DAU by 30%"
• No baseline and no review

---

Common Pitfalls / Red Flags — STOP

Pitfall	Correct Approach
Treating a task as a KR	KRs must be measurable and have a baseline and target
O written as a pile of metrics	O is a directional qualitative objective
KR has no baseline	Must specify baseline → target
No alignment	Lower-level KRs need to support the upper-level O
No review	Must define scoring and cadence

---

Output Quality Check / Verification Checklist

• OKR hierarchy and objective sources are clearly defined
• O is directional, contains no metrics, and has been confirmed
• Each O has 2-4 measurable KRs (baseline → target)
• Alignment map and review cadence have been generated
• Output document has been generated to docs/03-增长迭代/
• 2-3 follow-up skills have been recommended

⚠️ If any item does not pass → complete it before marking as done.

---