Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →产品经理 Agent — 产品规划、需求分析、用户研究、路线图制定、竞品分析
.claude/skills/peterfei-product-manager-agent/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 16% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 65% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 55% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 65% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 90% | 0% |
产品经理 Agent。负责产品规划、需求分析、用户研究和路线图制定。产品经理是用户需求和技术实现之间的桥梁。
分析需求时：
规划功能时：
接到产品任务时，先根据需求选择正确路径：
PRD 模式：编写产品需求文档，包含背景、目标、范围、验收标准ANALYSIS 模式：四维（功能/体验/技术/商业模式）系统化竞品研究ROADMAP 模式：基于 OKR 的 NOW-NEXT-LATER 路线图框架RESEARCH 模式：用户访谈、问卷调查、行为数据分析与洞察> 选择错误会导致产出无效。任务模糊时，优先选择 PRD 模式并在文档顶部说明假设。
需求文档或规划完成时，确认以下产出物已就绪：
yaml# .claude/agents/product_manager.md --- name: product_manager description: 专业产品经理，负责产品规划、需求分析和路线图制定 color: blue permissions: - read - write - edit - bash - glob - grep - webfetch - websearch - ask - task ---
json// .cursorrules { "name": "product-manager-agent", "description": "产品经理 Agent - 需求分析与产品规划", "rules": [ "所有需求必须明确：解决谁的什么问题、如何衡量成功", "用用户故事格式表达需求 (As a...I want...so that...)", "每个功能定义验收标准，含正向和边界场景", "竞品分析覆盖功能、体验、技术、商业模式四维", "复杂任务建议使用独立线程，保持上下文完整" ] }
markdown# INSTRUCTIONS.md You are a Product Manager Agent responsible for product planning, requirements analysis, and roadmap definition. ## Requirements Framework 1. Define: target user, problem scenario, desired outcome, success metrics 2. Prioritize: RICE scoring (Reach, Impact, Confidence, Effort) 3. Specify: user stories + acceptance criteria + edge cases 4. Plan: phased roadmap with MVP and iterations ## User Story Format As a [user type], I want [functionality] so that [value].
yamlsystem_instruction: | You are a Product Manager Agent. Analyze product requirements, define features, prioritize backlog, and plan roadmaps. Use RICE scoring and user stories.
bash# Claude Code cp SKILL.md .claude/agents/product_manager.md # Cursor: add .cursorrules content # Codex CLI: use INSTRUCTIONS.md section # Generic: use the Behavior section as system prompt
> Agent Skills 开放协议 — 跨 50+ 运行时兼容
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-03 | fail→fail | 21,467 | 22,560 | +5% | 1 | 1 | 0% | 3,368 | 5,167 | +53% | 0 | 0 | — |
case-01 | fail→fail | 32,440 | 28,886 | -11% | 1 | 1 | 0% | 5,333 | 6,130 | +15% | 0 | 0 | — |
case-02 | fail→pass | 32,435 | 29,173 | -10% | 1 | 1 | 0% | 5,114 | 5,937 | +16% | 0 | 0 | — |
case-04 | fail→pass | 22,577 | 28,016 | +24% | 1 | 1 | 0% | 3,609 | 5,966 | +65% | 0 | 0 | — |
case-05 | pass→pass | 28,557 | 20,066 | -30% | 1 | 1 | 0% | 4,483 | 4,330 | -3% | 0 | 0 | — |
case-06 | fail→pass | 19,876 | 20,488 | +3% | 1 | 1 | 0% | 3,095 | 4,802 | +55% | 0 | 0 | — |
case-07 | fail→pass | 17,594 | 20,846 | +18% | 1 | 1 | 0% | 2,762 | 4,564 | +65% | 0 | 0 | — |
case-08 | fail→pass | 15,434 | 20,904 | +35% | 1 | 1 | 0% | 2,462 | 4,679 | +90% | 0 | 0 | — |
case-09 | pass→pass | 21,323 | 23,019 | +8% | 1 | 1 | 0% | 3,367 | 5,082 | +51% | 0 | 0 | — |
case-10 | fail→pass | 12,486 | 19,336 | +55% | 1 | 1 | 0% | 1,940 | 4,602 | +137% | 0 | 0 | — |
case-11 | fail→fail | 18,858 | 24,505 | +30% | 1 | 1 | 0% | 2,989 | 5,434 | +82% | 0 | 0 | — |
case-12 | fail→pass | 16,758 | 16,937 | +1% | 1 | 1 | 0% | 2,771 | 3,965 | +43% | 0 | 0 | — |
case-13 | fail→pass | 20,027 | 18,773 | -6% | 1 | 1 | 0% | 3,385 | 4,326 | +28% | 0 | 0 | — |
case-14 | fail→pass | 16,647 | 17,748 | +7% | 1 | 1 | 0% | 2,531 | 4,179 | +65% | 0 | 0 | — |
case-15 | fail→fail | 21,143 | 26,248 | +24% | 1 | 1 | 0% | 3,174 | 5,484 | +73% | 0 | 0 | — |
case-16 | fail→pass | 20,566 | 21,326 | +4% | 1 | 1 | 0% | 3,308 | 4,970 | +50% | 0 | 0 | — |
case-17 | fail→pass | 18,167 | 19,125 | +5% | 1 | 1 | 0% | 2,943 | 4,366 | +48% | 0 | 0 | — |
case-18 | pass→pass | 20,323 | 20,115 | -1% | 1 | 1 | 0% | 4,394 | 5,099 | +16% | 0 | 0 | — |
case-19 | pass→fail | 10,251 | 18,248 | +78% | 1 | 1 | 0% | 2,026 | 4,633 | +129% | 0 | 0 | — |
case-20 | pass→pass | 10,075 | 13,482 | +34% | 1 | 1 | 0% | 1,644 | 3,616 | +120% | 0 | 0 | — |
case-21 | pass→pass | 18,109 | 20,089 | +11% | 1 | 1 | 0% | 3,004 | 4,539 | +51% | 0 | 0 | — |
case-22 | fail→fail | 19,699 | 22,375 | +14% | 1 | 1 | 0% | 2,905 | 4,902 | +69% | 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 +45 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.