When you need an AI workflow without building it from scratch

If you're trying to ship something like lead scoring, churn alerts, or a content pipeline without months of custom agent code, this repo is a cookbook of 36 recipes (skill chains) and 764 skills you can copy and adapt. Pre-built workflows and skills for Cursor and Claude; install via npm, pick a recipe, and wire your data. New to skills? What are skills?

Browse Skill Chain Recipes • Playbooks & Data to Wire • Explore the Directory • View Metrics • Quick Start

Choose your path

Guides by role — pick the one that matches your situation.

| Role | When this fits you | Get started | | :------------------ | :---------------------------------------------------------------------- | :-------------------------------------------- | | 👔 Sales Leader | You own pipeline and want AI qualification without a big build | View guide → | | 🚀 Growth PM | You're improving activation and want flows for onboarding or engagement | View guide → | | 🔬 Researcher | You need scientific tools and data analysis in your workflow | View guide → | | 💼 Finance/CFO | You need financial intelligence and reporting, not custom models | View guide → |

Or browse the Skill Chain Cookbook (36 recipes) → · browse 764 skills →

When this helps

You want a multi-step workflow (e.g. qualify → score → route) but don't want to design and maintain each step yourself. The cookbook gives you recipes; you copy one, point it at your CRM or data, and adjust. Skills are the building blocks; chains are the workflows that combine them.

• Single steps vs workflows: If you only need one-off prompts or single skills, you can use the skills library. If you need a repeatable flow (e.g. lead in → score → route → notify), the recipes are for that.
• Starting from zero vs from a recipe: Many teams build custom agents over several months. Here you start from a recipe and ship in days or weeks by wiring your data and tweaking prompts.
• Generic vs domain: Recipes are organized by domain (sales, churn, PLG, compliance, growth, etc.) so you can find something that matches what you're trying to do.

Example outcomes: Teams use the sales recipe to automate lead scoring and routing (fewer manual touches). Finance recipes are used for burn and scenario modeling. Churn recipes flag at-risk accounts early. Growth recipes drive conversion and onboarding flows. Full breakdown and numbers →

Installation

npm install @skene/skills-directory

Auto-activation

Skills are installed to your environment during npm install and work with:

• Cursor: ~/.cursor/skills/
• Claude: ~/.claude/skills/

Install is skipped automatically in:

• CI/CD (e.g. GitHub Actions, CircleCI, Jenkins)
• Docker
• When you use --ignore-scripts

To skip manually (e.g. in CI or Docker):

SKIP_SKILLS_INSTALL=true npm install @skene/skills-directory

An ECOSYSTEM.md file is generated with suggestions for other Skene tools that fit your stack.

What's included

If you'd rather copy a recipe and plug in your CRM/Stripe/playbook than build from scratch, start with the Skill Chain Cookbook (36 recipes). Each recipe chains 2–7 skills and is backed by a playbook (workflow blueprint) with ICP, integration references, and prompts — see Playbooks and registry/blueprints.

• 70 PLG: Product-led growth, activation, onboarding.
• 52 Marketing: Content, SEO, campaigns, analytics.
• 29 Customer Success: Health scoring, churn prediction.
• 25 RevOps: Sales pipeline, forecasting, GTM alignment.
• 20 Monetization: Pricing, billing, revenue optimization.
• 19 AI Ops: Automation and ML operations.
• 187 more: Security, data ops, devex, finance, HR, and others.

Executable skills breakdown →

• 241 Cursor rules: IDE guidelines for 241+ frameworks.
• 141 Scientific: Research tools, bioinformatics, data analysis.
• Integrations: 40+ standardized tools (CRM, analytics, billing).
• Security: Approval gates, rollbacks, audit trails.

• Blueprints: Each recipe maps to registry/blueprints/ with optional ICP, integration references, and per-step prompts.
• Data to wire: registry/integration_schemas/ has reference schemas for Salesforce, HubSpot, Stripe so you can wire CRM and billing without reinventing. See Playbooks.

Quick start

1. Install (see above).

2. See a working flow quickly — pick one by time you have:

   • ⏱️ 15 min: Lead scoring agent (2 skills)
   • 🕐 1 hour: Churn prevention agent (3 skills)
   • 🌅 Half day: Campaign automation agent (5 skills)

3. Check installation:

   npx skills-directory status

4. Deploy from a recipe: Browse the Skill Chain Cookbook — 36 recipes with step-by-step instructions across 15+ domains.

CLI reference

| Command | Description | | ------------------------------------------- | -------------------------------------- | | npx skills-directory status | Verify installation and file integrity | | npx skills-directory list --domain plg | List skills in a domain | | npx skills-directory ecosystem | Generate tool recommendations | | npx skills-directory showcase | Show what you can build (use cases) | | npx skills-directory install --target all | Manually install to Cursor + Claude | | npx skills-directory stats | Library statistics | | npx skills-directory uninstall | Remove installed skills |

Use cases (by situation)

Sales & RevOps

You already use a CRM and want leads scored and routed without writing the pipeline yourself. Recipes cover qualification, deal health, and forecasting.

Customer Success

You want to spot churn risk and trigger playbooks instead of building models from scratch. Recipes cover health monitoring, at-risk accounts (60–90 days early), and expansion/upsell triggers.

Finance & FinOps

You need burn, scenarios, and board-ready numbers without building everything in-house. Recipes cover ARR, burn rate, investor metrics, scenario planning, and board reporting.

Marketing & Growth

You want content or experiments wired into a workflow. Recipes cover content creation and distribution, conversion optimization, and programmatic SEO.

Detailed outcomes and calculations →

Documentation

Strategy and value

• VALUE.md — Outcomes and calculations for core use cases.
• SKILL_CHAINS.md — Skill Chain Cookbook: 36 recipes.
• SHOWCASE.md — Case studies and examples.
• QUICK_WINS.md — First agent in 15 min to 4 hours.

Technical and development

• Architecture — How the system works.
• Build Your First Skill — Create a skill step-by-step.
• Troubleshooting — Peer dependency warnings and common issues.
• Metrics methodology — Skill counts (764 total: 382 executable + 382 reference guides).
• Skills directory — Browse all skills by domain.
• Skill tree — Skills by domain (visual).
• AI Agent composability — Agent composition.

Contributing

To contribute to skene-cookbook, set up your environment:

git clone https://github.com/SkeneTechnologies/skene-cookbook.git
cd skene-cookbook

pip install -r requirements-test.txt
npm install
pre-commit install
pre-commit run --all-files

Pre-commit runs formatting, linting, secret checks. Bypass in emergencies: git commit --no-verify.

Before pushing to the public remote, run: ./scripts/pre_release_check.sh or npm run preflight. See CONTRIBUTING.md.

Community and support

Bug or question: Open an issue →

Contributing: Build Your First Skill · CONTRIBUTING.md · Code of Conduct

Security: Security Policy

Skill sources

Skills curated from: Anthropic Official, Trail of Bits, obra/superpowers, K-Dense-AI, awesome-cursor-rules, Skene PLG