shape-up-analysis-skills

内容来源:README.md(说明文档) · 原始地址 · 查看安装指南

原始内容

Shape Up Analysis Skills

Agent skills for analyzing Shape Up pitch documents alongside codebases to identify risks, rabbit holes, and unknowns before project kickoff.

What This Does

This skill transforms Claude into a specialized analyst that reviews Shape Up pitches with dual PM and Developer perspectives:

  • PM View: Validates problem clarity, success criteria, and business value
  • Dev View: Maps pitch requirements to actual code, identifies hidden complexity
  • Risk Analysis: Categorizes known risks, rabbit holes, and unknowns with mitigations
  • Actionable Output: Produces comprehensive kickoff document with confidence rating

Available Skills

Skill Description
shape-up-analysis Comprehensive pitch & codebase analysis for Shape Up projects

Installation

add-skill is a CLI tool that lets you install agent skills from git repositories. Works with Claude Code, OpenCode, Codex, Cursor, and more.

npx skills add tachun/shape-up-analysis-skills

or

npx add-skill tachun/shape-up-analysis-skills

Alternatively, manually copy skills/shape-up-analysis/ to your .claude/skills/ directory.

Usage

Once installed, trigger the skill by asking Claude to:

  • "Analyze this Shape Up pitch"
  • "Review this pitch against the codebase"
  • "Identify risks and rabbit holes in this pitch"
  • "Prepare a kickoff analysis"
  • "Assess this bet"
  • "Evaluate this pitch for the betting table"

Provide your pitch document and Claude will analyze it against the current codebase.

Output

The skill generates a project-kickoff-analysis.md with:

  1. Project Summary - TL;DR of problem, timing, and beneficiaries
  2. Problem Understanding - PM perspective on problem clarity and success criteria
  3. Solution Critical Review - Scope boundaries and assumption assessment
  4. Risks & Rabbit Holes - Categorized risks with likelihood, impact, and mitigations
  5. Codebase Review - Affected code areas with confidence levels and health signals
  6. Constraints & Boundaries - Time box and technical constraints
  7. Open Questions - Blocking questions with proceed/risk assessment
  8. Bet Confidence - 1-5 star rating with supporting factors
  9. Suggested Next Steps - Timeboxed spikes, experiments, and decisions needed
  10. Final Notes - Uncomfortable truths and accepted trade-offs

Repository Structure

shape-up-analysis-skills/
├── CLAUDE.md                 # Repository guide
├── AGENTS.md                 # Detailed agent instructions
├── README.md                 # This file
├── LICENSE                   # MIT license
└── skills/
    └── shape-up-analysis/
        ├── SKILL.md          # Main skill definition
        ├── references/
        │   ├── shape-up-methodology.md   # Shape Up concepts & terminology
        │   ├── risk-assessment.md        # Risk frameworks & scoring
        │   └── codebase-analysis.md      # Code review approach
        └── examples/
            ├── example-pitch.md          # Sample pitch document
            └── example-output.md         # Sample analysis output

Confidence Rating Scale

Rating Meaning Guidance
⭐⭐⭐⭐⭐ High confidence Green light, execute
⭐⭐⭐⭐☆ Good confidence Proceed with monitoring
⭐⭐⭐☆☆ Moderate confidence Address key risks first
⭐⭐☆☆☆ Low confidence Needs more shaping or spikes
⭐☆☆☆☆ Very low confidence Not ready to bet

About Shape Up

Shape Up is Basecamp's product development methodology featuring:

  • Fixed time, variable scope - Appetite sets the budget, scope flexes to fit
  • Pitches - Shaped proposals with problem, solution, rabbit holes, and no-gos
  • Betting table - Leadership decides which pitches get cycles
  • Six-week cycles - Big batches of focused work
  • Circuit breaker - Projects that don't converge get cut, not extended

This skill helps teams validate pitches before committing development time, reducing wasted effort and improving bet quality.

Contributing

  1. Fork the repository
  2. Create skill under skills/{skill-name}/SKILL.md
  3. Keep SKILL.md under 500 lines; use references/ for details
  4. Include examples demonstrating expected quality
  5. Submit pull request

License

MIT