Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Comprehensive toolkit for product managers including RICE prioritization, customer interview analysis, PRD templates, discovery frameworks, and go-to-market strategies. Use for feature prioritization, user research synthesis, requirement documentation, and product strategy development.
.claude/skills/anbeime-product-manager-toolkit/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 354% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 186% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 530% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 753% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 281% | 0% |
Essential tools and frameworks for modern product management, from discovery to delivery.
bashpython scripts/rice_prioritizer.py sample # Create sample CSV python scripts/rice_prioritizer.py sample_features.csv --capacity 15
bashpython scripts/customer_interview_analyzer.py interview_transcript.txt
references/prd_templates.mdbash # Create CSV with: name,reach,impact,confidence,effort python scripts/rice_prioritizer.py features.csv
bash python scripts/customer_interview_analyzer.py transcript.txt Extracts:
Advanced RICE framework implementation with portfolio analysis.
Features:
Usage Examples:
bash# Basic prioritization python scripts/rice_prioritizer.py features.csv # With custom team capacity (person-months per quarter) python scripts/rice_prioritizer.py features.csv --capacity 20 # Output as JSON for integration python scripts/rice_prioritizer.py features.csv --output json
NLP-based interview analysis for extracting actionable insights.
Capabilities:
Usage Examples:
bash# Analyze single interview python scripts/customer_interview_analyzer.py interview.txt # Output as JSON for aggregation python scripts/customer_interview_analyzer.py interview.txt json
Multiple PRD formats for different contexts:
Score = (Reach × Impact × Confidence) / Effort
Reach: # of users/quarter
Impact:
- Massive = 3x
- High = 2x
- Medium = 1x
- Low = 0.5x
- Minimal = 0.25x
Confidence:
- High = 100%
- Medium = 80%
- Low = 50%
Effort: Person-months Low Effort High Effort
High QUICK WINS BIG BETS
Value [Prioritize] [Strategic]
Low FILL-INS TIME SINKS
Value [Maybe] [Avoid]1. Context Questions (5 min)
- Role and responsibilities
- Current workflow
- Tools used
2. Problem Exploration (15 min)
- Pain points
- Frequency and impact
- Current workarounds
3. Solution Validation (10 min)
- Reaction to concepts
- Value perception
- Willingness to pay
4. Wrap-up (5 min)
- Other thoughts
- Referrals
- Follow-up permissionWe believe that [building this feature]
For [these users]
Will [achieve this outcome]
We'll know we're right when [metric]Outcome
├── Opportunity 1
│ ├── Solution A
│ └── Solution B
└── Opportunity 2
├── Solution C
└── Solution DAcquisition → Activation → Retention → Revenue → Referral
Key Metrics:
- Conversion rate at each step
- Drop-off points
- Time between steps
- Cohort variationsThis toolkit integrates with:
bash# Prioritization python scripts/rice_prioritizer.py features.csv --capacity 15 # Interview Analysis python scripts/customer_interview_analyzer.py interview.txt # Create sample data python scripts/rice_prioritizer.py sample # JSON outputs for integration python scripts/rice_prioritizer.py features.csv --output json python scripts/customer_interview_analyzer.py interview.txt json
本工具包可以与智能体协作框架无缝集成，实现产品团队的智能化协作。
将RICE排序工具与多智能体会议决策结合：
bash# 1. 使用RICE脚本生成初步排序 python scripts/rice_prioritizer.py features.csv --capacity 15 # 2. 调用智能体团队进行会议讨论 "请用产品团队评审以下功能的优先级：[功能列表]" # 3. 输出包含RICE分数和会议共识的完整决策
参与智能体:
相关技能:
agent-team: 智能体协作框架multi-agent-meeting: 会议决策流程会议模板: multi-agent-meeting/assets/meeting-templates/product-feature-review.md
将访谈分析工具与智能体团队结合：
bash# 1. 使用分析脚本提取洞察 python scripts/customer_interview_analyzer.py interview.txt # 2. 调用智能体团队讨论改进方案 "分析这份访谈记录并生成产品改进方案：[访谈文本]" # 3. 输出包含洞察、方案和PRD草稿的完整报告
参与智能体:
相关技能:
agent-team: 智能体协作框架multi-agent-meeting: 会议决策流程会议模板: multi-agent-meeting/assets/meeting-templates/customer-insight-analysis.md
结合战略分析和RICE排序：
bash# 1. 战略分析智能体识别市场机会 "扫描AI内容生成领域的市场机会" # 2. 使用RICE方法对机会进行排序 python scripts/rice_prioritizer.py opportunities.csv --capacity 15 # 3. 召开路线图规划会议 "制定Q2产品路线图，团队容量15人月" # 4. 输出季度路线图和资源分配计划
参与智能体:
相关技能:
agent-team: 智能体协作框架（场景9：产品路线图规划会议）当在智能体协作框架中使用"产品经理智能体"时，该智能体具备以下能力：
专业知识:
工具调用:
python# 优先级排序 self.call_tool("rice_prioritizer", features_csv, capacity=15) # 访谈分析 self.call_tool("customer_interview_analyzer", interview_text) # PRD生成 self.use_template("prd_templates", template_type="standard")
协作接口:
适用场景:
智能体定义: 参考 agent-team/references/agent-registry.md 中的"产品经理智能体（增强版）"
专业知识:
工具调用:
python# 访谈分析 self.call_tool("customer_interview_analyzer", interview_text, output_format="json") # 提取关键洞察 insights = self.extract_insights(analysis_result)
协作接口:
适用场景:
智能体定义: 参考 agent-team/references/agent-registry.md 中的"用户研究员智能体"
用户输入:
我有以下功能需要评审优先级：
1. AI自动生成脚本
2. 多平台发布
3. 数据分析看板
4. 用户协作功能
5. API开放平台
请帮我评审并生成优先级排序，团队容量15人月。系统执行流程:
bash # 产品经理智能体调用RICE脚本 python scripts/rice_prioritizer.py features.csv --capacity 15
> 产品经理: 根据RICE排序，AI自动生成脚本得分最高(85分)... > 技术架构师: 从技术角度，这个功能可行，可以使用现有大模型... > 市场分析师: 市场调研显示，用户对AI功能的需求强烈... > 财务顾问: 预计3个月可以回本，ROI较高...
markdown # 功能优先级排序结果
## 最终排序
## Q2开发计划
智能体协作框架:
agent-teamagent-team/references/agent-registry.mdagent-team/references/collaboration-templates.md会议决策流程:
multi-agent-meetingmulti-agent-meeting/assets/meeting-templates/multi-agent-meeting/references/meeting-record-format.md产品场景模板:
整合前:
整合后:
用户价值:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 5,657 | 6,795 | +20% | 1 | 1 | 0% | 1,145 | 5,197 | +354% | 0 | 0 | — |
case-02 | pass→pass | 6,179 | 4,777 | -23% | 1 | 1 | 0% | 933 | 4,683 | +402% | 0 | 0 | — |
case-03 | pass→pass | 3,173 | 2,383 | -25% | 1 | 1 | 0% | 487 | 4,387 | +801% | 0 | 0 | — |
case-04 | pass→pass | 5,190 | 3,365 | -35% | 1 | 1 | 0% | 1,039 | 4,605 | +343% | 0 | 0 | — |
case-05 | pass→pass | 4,637 | 12,406 | +168% | 1 | 1 | 0% | 869 | 4,310 | +396% | 0 | 0 | — |
case-06 | fail→pass | 7,983 | 1,836 | -77% | 1 | 1 | 0% | 1,496 | 4,277 | +186% | 0 | 0 | — |
case-07 | fail→pass | 4,007 | 3,425 | -15% | 1 | 1 | 0% | 720 | 4,533 | +530% | 0 | 0 | — |
case-08 | pass→pass | 4,203 | 2,034 | -52% | 1 | 1 | 0% | 776 | 4,322 | +457% | 0 | 0 | — |
case-09 | pass→pass | 3,443 | 13,189 | +283% | 1 | 1 | 0% | 668 | 4,407 | +560% | 0 | 0 | — |
case-10 | fail→pass | 3,443 | 4,190 | +22% | 1 | 1 | 0% | 532 | 4,538 | +753% | 0 | 0 | — |
case-21 | pass→pass | 9,275 | 9,914 | +7% | 1 | 1 | 0% | 2,168 | 6,289 | +190% | 0 | 0 | — |
case-11 | fail→pass | 6,719 | 2,135 | -68% | 1 | 1 | 0% | 1,138 | 4,341 | +281% | 0 | 0 | — |
case-12 | pass→pass | 4,164 | 8,626 | +107% | 1 | 1 | 0% | 659 | 4,414 | +570% | 0 | 0 | — |
case-13 | pass→pass | 4,841 | 2,152 | -56% | 1 | 1 | 0% | 886 | 4,310 | +386% | 0 | 0 | — |
case-14 | fail→pass | 5,166 | 1,949 | -62% | 1 | 1 | 0% | 968 | 4,289 | +343% | 0 | 0 | — |
case-19 | pass→pass | 3,745 | 2,504 | -33% | 1 | 1 | 0% | 745 | 4,330 | +481% | 0 | 0 | — |
case-15 | fail→pass | 7,515 | 1,657 | -78% | 1 | 1 | 0% | 1,295 | 4,171 | +222% | 0 | 0 | — |
case-16 | pass→pass | 2,207 | 2,826 | +28% | 1 | 1 | 0% | 369 | 4,438 | +1103% | 0 | 0 | — |
case-17 | fail→pass | 8,509 | 3,074 | -64% | 1 | 1 | 0% | 1,343 | 4,549 | +239% | 0 | 0 | — |
case-18 | fail→pass | 8,399 | 2,191 | -74% | 1 | 1 | 0% | 1,128 | 4,287 | +280% | 0 | 0 | — |
case-20 | pass→pass | 11,590 | 12,018 | +4% | 1 | 1 | 0% | 2,498 | 6,428 | +157% | 0 | 0 | — |
case-22 | pass→pass | 14,951 | 14,733 | -1% | 1 | 1 | 0% | 2,129 | 6,607 | +210% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +41 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.