Skip to content

ashstep2/pm-playbook

BranchesTags

Repository files navigation

PM Playbooks

30 reusable PM skills for Claude Code — repeatable decision-making workflows that produce structured artifacts (scorecards, frameworks, roadmaps) grounded in attributed insights from product leaders.

Skills Meta-skills Collectors License

Architecture

┌─────────────────────────────────────────────────────────────┐
│                     Your Company Repo                       │
│                                                             │
│  context/            .claude/skills/           applied/     │
│  ├── company/        ├── customer-discovery    ├── strategic│
│  ├── competitive/    ├── feature-prioritization├── planning │
│  ├── products/       ├── north-star-metrics    └── ...      │
│  ├── verticals/      ├── ... (30 skills)                    │
│  └── signals/        └── ... (+ company skills)             │
│       ▲                     │                      │        │
│       │                     ▼                      │        │
│  collectors/         skill-graph.yaml              │        │
│  (auto-refresh)      (dependency map)              │        │
│                                                             │
│  pm-playbooks/  ←── git submodule ──────────────────        │
└─────────────────────────────────────────────────────────────┘

Quick Start

git clone https://github.com/ashstep2/pm-playbook.git
cd pm-playbooks
# Open in Claude Code and invoke any skill by name

No configuration required. Skills work out of the box by asking for context interactively. Add a context/ directory to ground outputs in your company's data.

New here? Start with the Getting Started Guide — pick a journey that matches your situation and run it. Or browse example outputs to see what skills produce.

Why This Is Different

Generic prompts Consulting frameworks pm-playbooks
Structured output Free-form text PDF decks Markdown artifacts with sections, tables, scorecards
Attributed principles "Best practice says..." Proprietary IP Named sources with direct quotes and URLs
Context-aware Starts from zero Manual input Reads context/ directory, fills gaps conversationally
Composable One-shot Siloed engagements Skills chain via skill-graph.yaml
Evolving Static Annual updates /learn ingests new content, /improve reflects on runs
Signal-fed No live data Expensive research Auto-collects from GitHub, HN, Reddit, news APIs

Skills

Strategic — Deciding WHAT to build

Skill What It Produces
product-portfolio-strategy Multi-product roadmap + resource allocation across bets
measuring-product-market-fit PMF scorecard + signal tracking + pivot/persevere framework
competitive-response Threat assessment + response playbook + moat analysis
research-to-product-pipeline Translation framework: research breakthroughs → shipped products
platform-vs-application Build/buy/partner analysis + platform economics + ecosystem design
vertical-market-assessment TAM/SAM/SOM + vertical prioritization matrix + entry strategy
product-narrative-strategy Product vision doc + strategy narrative + roadmap story + customer context brief

Planning — Deciding HOW to build

Skill What It Produces
zero-to-one-product-launch Launch checklist + channel strategy + success metrics + rollback plan
feature-prioritization Weighted scoring matrix + sequenced roadmap + trade-off analysis
pricing-and-monetization Pricing model analysis + willingness-to-pay + packaging strategy
north-star-metrics Metric tree + leading/lagging indicators + dashboard design
writing-prds-for-ai PRD template + uncertainty handling + eval criteria for AI outputs
experiment-design Experiment briefs + statistical plans + growth loop design + results synthesis

Execution — Building and shipping

Skill What It Produces
customer-discovery Interview guide + synthesis framework + discovery report
developer-experience-audit Friction scorecard + prioritized DX recommendations
user-onboarding-optimization Onboarding flow audit + activation metrics + improvement roadmap
stakeholder-alignment RACI matrix + communication plan + research-product interface
prototype-driven-validation Prototype brief + variation matrix + customer test plan + build/kill decision
iteration-cadence-design Cadence architecture + weekly PM calendar + ritual design + adaptation triggers

Agent-First — Building for agent-mediated distribution

Skill What It Produces Persona
agent-surface-audit Agent accessibility scorecard + surface inventory + remediation roadmap Developer / All
mcp-design-review Per-tool scorecard + security audit + description quality assessment Developer
agent-journey-mapping Agent journey maps + I/O contracts + failure mode analysis Developer / Ops
agent-ready-gtm Agent buyer journey + trust center API + pricing surface design Enterprise Buyer
agent-integration-design I/O specs + webhook design + orchestration compatibility matrix Ops / Workflow
agent-consumer-experience Agent interaction map + consent framework + preference API End User

Analysis — Measuring and learning

Skill What It Produces
user-segmentation Segment profiles + prioritization matrix + cross-segment insights
ecosystem-health Ecosystem scorecard + benchmark comparison + growth playbook
api-design-review Per-endpoint assessment + standards compliance + agent readiness
partnership-evaluation Partner scorecards + deal structures + comparative assessment
product-quality-review Quality scorecard + taste gap analysis + competitive craft comparison + quality roadmap

Skill Dependency Graph

Skills feed into each other. After completing a skill, the system suggests what to run next based on skill-graph.yaml.

graph LR
    subgraph Strategic
        PPS[product-portfolio-strategy]
        PMF[measuring-product-market-fit]
        CR[competitive-response]
        RTP[research-to-product-pipeline]
        PVA[platform-vs-application]
        VMA[vertical-market-assessment]
        PNS[product-narrative-strategy]
    end

    subgraph Planning
        ZTO[zero-to-one-product-launch]
        FP[feature-prioritization]
        PM[pricing-and-monetization]
        NSM[north-star-metrics]
        PRD[writing-prds-for-ai]
        EXP[experiment-design]
    end

    subgraph Execution
        CD[customer-discovery]
        DXA[developer-experience-audit]
        UOO[user-onboarding-optimization]
        SA[stakeholder-alignment]
        PDV[prototype-driven-validation]
        ICD[iteration-cadence-design]
    end

    subgraph Agent-First
        ASA[agent-surface-audit]
        MDR[mcp-design-review]
        AJM[agent-journey-mapping]
        AGTM[agent-ready-gtm]
        AID[agent-integration-design]
        ACE[agent-consumer-experience]
    end

    subgraph Analysis
        US[user-segmentation]
        EH[ecosystem-health]
        ADR[api-design-review]
        PE[partnership-evaluation]
        PQR[product-quality-review]
    end

    CD --> US --> PMF --> ZTO
    CR --> FP
    CR --> PPS
    CR --> ASA
    EH --> CR
    EH --> PE
    EH --> DXA
    NSM --> FP
    PPS --> FP
    PPS --> RTP
    FP --> PRD --> ZTO
    FP --> ZTO
    ZTO --> UOO
    ZTO --> SA
    VMA --> CD
    DXA --> ADR
    DXA --> ASA
    ADR --> MDR
    ADR --> ASA
    PE --> PPS
    RTP --> PRD
    ASA --> MDR
    ASA --> AJM
    ASA --> AGTM
    MDR --> AJM
    MDR --> AID
    AJM --> AID
    AJM --> ACE
    AGTM --> ZTO
    ACE --> UOO
    CD --> PDV
    FP --> PDV
    PDV --> PRD
    PDV --> SA
    PDV --> ZTO
    PDV --> EXP
    PNS --> SA
    PNS --> ZTO
    PNS --> FP
    CD --> PNS
    CR --> PNS
    NSM --> EXP
    EXP --> PMF
    EXP --> UOO
    ICD --> PDV
    ICD --> EXP
    DXA --> PQR
    PQR --> FP
    PQR --> PDV
Loading

Installation

Path A: Zero-Config (30 seconds)

git clone https://github.com/ashstep2/pm-playbook.git
# Open in Claude Code → invoke any skill by name

Skills ask for context interactively. No setup required.

Path B: Project Integration (5 minutes)

# From your project root
git submodule add https://github.com/ashstep2/pm-playbook.git
bash pm-playbooks/scaffold/install.sh

This symlinks all 30 skills into .claude/skills/ and generates a CLAUDE.md for your project.

Path C: Full Company Setup (30 minutes)

git submodule add https://github.com/ashstep2/pm-playbook.git
bash pm-playbooks/scaffold/install.sh

# Fill in company context
mkdir -p context/{company,competitive,products,verticals,founders,signals}
# Add markdown files to each directory (see scaffold/ for templates)

# Set up signal collectors (optional)
cp pm-playbooks/signals.yaml.example signals.yaml
# Edit signals.yaml, add API keys to .env
pip install pyyaml requests PyGithub praw
cd pm-playbooks && python3 -m collectors.run --config ../signals.yaml

Graceful Degradation

Skills work at every level — more context produces more grounded output, but nothing breaks without it.

Setup Level Behavior
No context/ Skills ask for information interactively
Partial context/ Use what's available, ask for the rest
Full context/ + signals Fully grounded, no questions needed

Self-Improvement

Two meta-skills in _meta/ help the system evolve:

  • /learn — Ingest an article, podcast transcript, or book excerpt. Extracts insights, filters for novelty, and writes improvement proposals to _meta/proposals/ for human review. Proposals are never auto-applied.
  • /improve — Run after any skill to score its effectiveness (Instruction Clarity, Context Sufficiency, Artifact Usefulness) and propose one concrete improvement. Reflections accumulate in applied/_reflections.md and surface recurring patterns.

See CONTRIBUTING.md for how proposals become PRs.

Signal Collectors

The collectors/ framework auto-collects public data into context/signals/.

Collector Auth Required What It Collects
hackernews None HN stories matching configured keywords
github GITHUB_TOKEN Repo stars, forks, issues, PRs for tracked repos
reddit REDDIT_CLIENT_ID + REDDIT_CLIENT_SECRET Posts from configured subreddits
news NEWS_API_KEY Articles from NewsAPI 
cd pm-playbooks && python3 -m collectors.run --config ../signals.yaml

Config lives in signals.yaml at your repo root. See signals.yaml.example for format.

Skill Format

Each skill follows the Anthropic Agent Skills spec. See SKILL_FORMAT.md for the full reference. Skills include:

  • Core Principles — attributed insights with direct quotes and source URLs
  • Instructions — step-by-step workflow producing concrete artifacts
  • Diagnostic Questions — assess the situation before diving in
  • Common Mistakes — what PMs get wrong and how to avoid it
  • Context Integration — how the skill uses company data

Contributing

See CONTRIBUTING.md for how to add or improve skills.

License

MIT

About

A "PM in your pocket." Decision-making workflows that produce structured artifacts. A step on the journey to automating the product role.

Resources

Readme

License

MIT license

Contributing

Contributing
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages