搜索

PythonReactSEOPDF
GitHub Skill

product-owner

Assist Product Owners with idea capture, backlog management, prioritization frameworks, and stakeholder communication. Use when working with product ideas, backlog items, sprint planning, or prioritization decisions. Triggers on keywords like "idea", "backlog", "prioritize", "sprint planning", "user value", "stakeholder", "product vision".

duc01226☆ 3 仓库 Stars1 仓库 Forks

安装与使用

请根据来源仓库和你的 AI 工具说明安装此技能。先阅读技能文件和依赖要求,再添加到工作流。

查看来源 ↗使用指南

技能说明

Product Owner Assistant

Help Product Owners capture ideas, manage backlogs, and make prioritization decisions using established frameworks.


Core Capabilities

1. Idea Capture
  • Transform raw concepts into structured idea artifacts
  • Identify problem statements and value propositions
  • Tag and categorize for future refinement
2. Backlog Management
  • Create and refine Product Backlog Items (PBIs)
  • Maintain backlog ordering (not categories)
  • Track dependencies and blockers
3. Prioritization Frameworks
RICE Score
RICE = (Reach × Impact × Confidence) / Effort

Reach: # users affected per quarter
Impact: 0.25 (minimal) | 0.5 (low) | 1 (medium) | 2 (high) | 3 (massive)
Confidence: 0.5 (low) | 0.8 (medium) | 1.0 (high)
Effort: Person-months
MoSCoW
  • Must Have: Critical for release, non-negotiable
  • Should Have: Important but not vital
  • Could Have: Nice to have, low effort
  • Won't Have: Out of scope this cycle
Value vs Effort Matrix
         High Value
             │
    Quick    │    Strategic
    Wins     │    Priorities
─────────────┼─────────────
    Fill     │    Time
    Ins      │    Sinks
             │
         Low Value
   Low Effort    High Effort
4. Sprint Planning Support
  • Capacity planning based on velocity
  • Sprint goal definition
  • Commitment vs forecast distinction

Artifact Templates

Idea Template Location

team-artifacts/templates/idea-template.md

PBI Template Location

team-artifacts/templates/pbi-template.md


Business Features Reference

Context Loading for Ideas

When creating ideas via /idea command, load relevant business documentation:

  1. Identify Module: Match idea keywords to docs/business-features/ modules
  2. Load INDEX.md: Get feature table for overlap detection
  3. Load README.md: Extract requirements (FR-XX) for reference
  4. Check Gaps: Note if idea fills documented gap vs extends existing
Module Documentation Structure
docs/business-features/{Module}/
├── INDEX.md          # Feature navigation table
├── README.md         # 15-section module documentation
├── API-REFERENCE.md  # Endpoint documentation
└── detailed-features/
    └── *.md          # Deep-dive feature docs
Dynamic Module Discovery

Modules are discovered from frontmatter in docs/business-features/*/README.md:

# Frontmatter schema (see docs/templates/detailed-feature-docs-template.md)
module: ModuleName
keywords: [term1, term2]      # Match against user input
aliases: [shortname]          # Exact match
features: [sub-feature]       # Sub-feature keywords
domain_path: src/.../Domain   # For entity inspection

Discovery Algorithm:

  1. Glob all module README.md files
  2. Parse YAML frontmatter
  3. Match user keywords against keywords, aliases, features
  4. Return module with highest match count
  5. Fallback: prompt user to select from available modules
Related Workflows
  • /idea command discovers modules via frontmatter and loads context
  • /refine uses domain_path from frontmatter for entity inspection
  • Use INDEX.md feature table to prevent duplicate work
  • Reference existing FR-XX IDs when related

Workflow Integration

Creating Ideas

When user says "new idea" or "feature request":

  1. Use /idea command workflow
  2. Populate idea-template.md
  3. Save to team-artifacts/ideas/
  4. Suggest next step: /refine {idea-file}
Prioritizing Backlog

When user says "prioritize" or "order backlog":

  1. Read all PBIs in team-artifacts/pbis/
  2. Apply requested framework (RICE, MoSCoW, Value/Effort)
  3. Output ordered list with scores
  4. Update priority field in PBI frontmatter

Output Conventions

File Naming
{YYMMDD}-po-idea-{slug}.md
{YYMMDD}-pbi-{slug}.md
Priority Values
  • Numeric ordering: 1 (highest) to 999 (lowest)
  • Never use High/Medium/Low categories
Status Values

draft | under_review | approved | rejected | in_progress | done


Anti-Patterns to Avoid

  1. Category-based priority - Use ordered sequence, not High/Med/Low
  2. Vague acceptance criteria - Require GIVEN/WHEN/THEN format
  3. Scope creep - Explicitly list "Out of Scope"
  4. Missing dependencies - Always identify upstream/downstream

Integration Points

When Trigger Action
Idea captured /idea complete Suggest /refine
PBI ready PBI approved Notify BA for stories
Sprint planned Sprint goal set Update PBI assignments

Stakeholder Communication Templates

Sprint Review Summary
## Sprint {N} Review

**Sprint Goal:** {goal}
**Status:** {achieved | partially | not achieved}

### Completed Items
| PBI | Value Delivered |
|-----|-----------------|
| | |

### Carried Over
| PBI | Reason | Plan |
|-----|--------|------|
| | | |

### Key Metrics
- Velocity: {points}
- Commitment: {%}
Roadmap Update
## Roadmap Update - {Date}

### This Quarter
| Priority | Item | Target | Status |
|----------|------|--------|--------|
| 1 | | | |

### Next Quarter
| Item | Dependencies | Notes |
|------|--------------|-------|
| | | |

### Deferred
| Item | Reason |
|------|--------|
| | |

Quality Checklist

Before completing PO artifacts:

  • Problem statement is user-focused, not solution-focused
  • Value proposition quantified or qualified
  • Priority has numeric order
  • Dependencies explicitly listed
  • Status frontmatter current

Task Planning Notes

  • Always plan and break many small todo tasks
  • Always add a final review todo task to review the works done at the end to find any fix or enhancement needed

使用示例

查看来源仓库中的使用示例与配置说明。

查看来源 ↗

来源历史

收录时间
2026-01-20
索引更新时间
2026-01-28
Branch
main

收录与同步日期属于本站记录。源码修改历史请查看来源仓库。

GitHub ↗

chat Comments (0)

chat_bubble_outline

No comments yet. Be the first to share your thoughts!