uthy-legacy-os
v1.0.2
Published
Product Management CLI — AI-powered PM skills, chat, config, workflows. 35 skills for discovery, PRDs, shipping, measuring.
Maintainers
Readme
╔══════════════════════════════════════╗
║ UTHY LEGACY OS — v1.0.0 ║
║ Product Management Skills for AI ║
╚══════════════════════════════════════╝
30 battle-tested PM skills + 4 guided workflows for Claude Code, Cursor, Codex, Gemini CLI, Copilot, and any AI agent that speaks SKILL.md.
"Skills provide expertise; workflows provide momentum."
What is Uthy Legacy OS?
A portable, open‑source library of product management skills that turns any AI coding agent into a capable PM partner. Each skill is a SKILL.md file — a standard format supported by Claude Code, Cursor, Codex CLI, Gemini CLI, Copilot, Windsurf, and 20+ other agents.
Install once. Use everywhere. Never start from scratch again.
What's Inside
| Phase | Skills | Covers | |-------|--------|--------| | 🔍 Discovery | 5 skills | Customer interviews, competitive analysis, market sizing, user research, opportunity trees | | 🎯 Problem Definition | 5 skills | Problem statements, JTBD, personas, framing, elevator pitches | | 💡 Solution Exploration | 5 skills | Brainstorming, hypothesis prioritization, experiment design, prototyping, validation | | 📝 PRDs & Stories | 5 skills | PRD writer, user stories, acceptance criteria, epic breakdown, spec bridges | | 🚀 Shipping | 5 skills | Launch plans, release notes, GTM, rollout strategy, stakeholder comms | | 📊 Measuring | 5 skills | Metrics frameworks, experiment analysis, dashboards, A/B tests, north stars | | ♻️ Improve | 5 skills | Retros, feedback analysis, iteration prioritization, churn, adoption funnels | | 🔄 Workflows | 4 workflows | Discovery sprint, ship cycle, quarterly planning, health review |
Quick Install
Choose your weapon — all four installers do the same thing.
🟢 npm (Node.js)
npx uthy-legacy-os
# or install globally once:
npm install -g uthy-legacy-os && uthy-legacy-os🔵 pnpm
pnpm dlx uthy-legacy-os🟣 bun
bunx uthy-legacy-os🟡 PowerShell (Windows native)
.\install.ps1🔴 Bash (Linux / macOS / Git Bash)
git clone https://github.com/uthumany/uthy-legacy-os.git ~/.uthy-legacy-os
cd ~/.uthy-legacy-os && bash install.shPer‑Agent Manual Install
If you prefer to symlink manually:
Claude Code
mkdir -p ~/.claude/skills
ln -s ~/.uthy-legacy-os/skills/*/* ~/.claude/skills/Cursor
mkdir -p ~/.cursor/skills
ln -s ~/.uthy-legacy-os/skills/*/* ~/.cursor/skills/Codex CLI
mkdir -p ~/.codex/skills
ln -s ~/.uthy-legacy-os/skills/*/* ~/.codex/skills/Hermes Agent
hermes curator install url https://github.com/uthumany/uthy-legacy-osGemini CLI
mkdir -p ~/.gemini/skills
ln -s ~/.uthy-legacy-os/skills/*/* ~/.gemini/skills/How It Works
Each skill is a folder containing:
skill-name/
├── SKILL.md # Instructions + YAML frontmatter (the core)
├── references/ # Deep-dive docs, frameworks, checklists
├── templates/ # Reusable templates the agent can fill in
└── samples/ # Realistic outputs that show "what good looks like"When you ask your agent to "write a PRD" or "plan a discovery sprint", the agent's skill system automatically finds the matching skill and loads it. The skill teaches the agent:
- What to do — step‑by‑step guidance
- How to do it — frameworks, templates, pro tips
- What good looks like — sample outputs and quality standards
- What to avoid — pitfalls and anti‑patterns
Skill Catalog (30 Skills)
| Skill | Phase | What It Does |
|-------|-------|-------------|
| customer-interviews | Discovery | Plan and conduct customer discovery interviews |
| competitive-analysis | Discovery | Deep‑dive competitive positioning analysis |
| market-sizing | Discovery | TAM/SAM/SOM with top‑down and bottom‑up |
| user-research | Discovery | Synthesis of qualitative and quantitative research |
| opportunity-solution-tree | Discovery | Map opportunities to potential solutions (Teresa Torres) |
| problem-statement | Problem | Frame the customer problem with evidence |
| jobs-to-be-done | Problem | Understand customer goals (JTBD framework) |
| proto-persona | Problem | Hypothesis‑driven personas before research |
| problem-framing | Problem | Reframe problems to unlock better solutions |
| elevator-pitch | Problem | One‑sentence value propositions (Geoffrey Moore) |
| solution-brainstorming | Solutions | Structured ideation with constraints |
| hypothesis-prioritization | Solutions | RICE, ICE, and value‑effort quadrants |
| experiment-design | Solutions | Design lightweight experiments (PoL method) |
| prototype-planning | Solutions | Plan what to prototype and at what fidelity |
| validation-interview | Solutions | Test prototypes with customer interviews |
| prd-writer | PRD & Stories | Write comprehensive PRDs from research |
| user-story-writer | PRD & Stories | Stories with acceptance criteria (Gherkin) |
| acceptance-criteria | PRD & Stories | Write clear, testable criteria (Given/When/Then) |
| epic-breakdown | PRD & Stories | Split epics into deliverable stories |
| technical-spec-bridge | PRD & Stories | Translate PRDs for engineering audiences |
| launch-plan | Shipping | End‑to‑end launch checklists and timing |
| release-notes | Shipping | Write release notes for different audiences |
| go-to-market | Shipping | GTM strategy for new features/products |
| rollout-strategy | Shipping | Phased rollout plans with rollback |
| stakeholder-communication | Shipping | Status updates, exec briefs, and dashboards |
| metrics-framework | Measuring | AARRR, HEART, and custom metric trees |
| experiment-analysis | Measuring | Analyze experiment data and draw conclusions |
| dashboard-design | Measuring | Design metric dashboards that drive action |
| ab-test-reader | Measuring | Interpret A/B test results with confidence |
| north-star-metrics | Measuring | Define and align around north star metrics |
| retrospective | Improve | Structured retros (Start/Stop/Continue, 4Ls) |
| feedback-analysis | Improve | Synthesize user feedback into insights |
| iteration-prioritization | Improve | Prioritise post‑launch improvements |
| churn-analysis | Improve | Diagnose churn and identify retention levers |
| adoption-funnel | Improve | Analyze and improve feature adoption |
Guided Workflows
Beyond individual skills, Uthy Legacy OS includes 4 cross‑phase workflows that orchestrate multiple skills together:
| Workflow | Duration | Skills Used |
|----------|----------|-------------|
| product-discovery-sprint | 1–2 weeks | 8 skills across discovery → problem → solutions |
| feature-ship-cycle | 2–4 weeks | 10 skills from PRD → shipping → measuring |
| quarterly-planning | 1 week | 6 skills for strategic planning |
| product-health-review | Weekly | 4 skills for ongoing health monitoring |
Each workflow is a SKILL.md that acts as a sub‑agent orchestrator, calling other skills in sequence with handoff criteria.
Project Structure
uthy-legacy-os/
├── README.md # You are here
├── AGENTS.md # Instructions for AI contributors
├── LICENSE # MIT
├── CONTRIBUTING.md # How to add a skill
├── install.sh # Bash installer (Linux/macOS/Git Bash)
├── install.ps1 # PowerShell installer (Windows native)
├── bin/cli.js # CLI entry point (npm/pnpm/bun)
├── package.json # npm package metadata
├── skills/ # 30 skills in 6 phases + 4 workflows
├── templates/ # Reusable document templates
├── samples/ # Realistic sample outputs
├── docs/ # Documentation site (GitHub Pages)
├── scripts/ # Validation and release tooling
└── .github/workflows/ # CI/CDDevelopment & Contribution
We welcome contributions! See CONTRIBUTING.md for:
- Skill authoring guide
- Quality standards
- Validation checks
- PR workflow
Before submitting, run the validator:
python scripts/validate-skills.pyLicense
MIT — free to use, fork, and share. Attribution appreciated but not required.
About
Built by Uthuman & Co for product managers who believe AI shouldn't just write code — it should help you think better about what to build.
