The skill library
Hiring an AI employee usually means getting something clever and completely uninformed, and then spending a month explaining your job to it. Ours arrive knowing 73 methods, the ones a good operator would have read over a career.
A skill is a written method an employee follows when the work calls for it. Ask one to price a new plan and it does not improvise: it works through the value equation, checks the result against the ways that usually go wrong, and tells you what it would change.
Skills come with the hire. You do not assemble them, and you are not writing prompts to fake them. You can add your own on top, and an employee can write one for you from a job you have described once.
Of the 73 below, 50 adapt frameworks from published books and are used under the MIT licence, with credit to the authors who wrote them. We name the method rather than reprint it, because the book is theirs and worth buying.
What you sell, what you charge, and why somebody should pick you over the alternative.
Make an offer irresistible: value equation, bonus stacks, risk-reversing guarantees, ethical scarcity, naming.
Design pricing and packaging around validated willingness to pay: tiers, freemium vs trial, price rises, good-better-best.
Position a product against its alternatives: unique attributes, best-fit customers, market category, repositioning.
Take a product from early adopters across the chasm to mainstream buyers: beachhead segment, whole product, pragmatists.
Find uncontested market space instead of competing on price: value innovation, eliminate-reduce-raise-create, non-customers, strategy canvas, buyer utility.
Create or update .agents/product-marketing.md, context other marketing skills read: product, ICP, audience, positioning.
Find the job a customer hires a product to do: switching behavior, functional and emotional jobs, why they pick a competitor.
Build a quiz, assessment, calculator, or grader as a lead magnet: question design, scored tiers, result pages, follow-up.
Saying it so it lands, in a landing page, an email, or a sentence somebody repeats to a colleague.
Clarify brand messaging with the StoryBrand SB7 structure: customer as hero, one-liner, homepage and email copy.
Write or rewrite marketing page copy: homepage, landing, pricing, feature, about. Headlines, value props, CTAs.
Edit, review, or refresh existing copy rather than rewrite it: tighten, proofread, sharpen messaging, update stale content.
Make a message stick with the SUCCESs checklist: taglines, value props, pitch decks, explaining complex ideas simply.
Engineer word of mouth and sharing: social currency, triggers, emotion, public visibility, practical value, stories.
Apply ethical persuasion to copy, product, and sales: reciprocity, commitment, social proof, authority, liking, scarcity.
Build an end-to-end marketing plan: target market, USP, channels, lead nurture, lifetime value, referrals, the PVP Index.
Design a multi-email automated flow: welcome, onboarding, nurture, re-engagement, post-purchase, lifecycle.
Write B2B cold outreach and follow-up sequences that get replies: subject lines, openers, CTAs, personalization.
Search, content, social and paid, including how to be cited by an answer engine rather than only ranked by a search engine.
Get content cited by AI search: AI Overviews, ChatGPT, Perplexity, Gemini. Covers AEO, GEO, LLMO.
Audit a site's SEO: lost rankings, traffic drop, indexing, on-page and technical issues, core web vitals.
Generate many templated SEO pages from data: location, comparison, integration, and directory pages, without thin content.
Add or fix schema.org structured data and JSON-LD for Google rich results: FAQ, product, review, breadcrumb.
Plan a site's page hierarchy, navigation, URL structure, internal linking, or a visual sitemap.
Decide what content to create and why: topics, pillars, clusters, editorial calendar, roadmap.
Create, schedule, or repurpose social posts and short-form video, or triage engagement. Any platform.
Plan, optimize, or scale paid ad campaigns on any platform: targeting, bidding, budget, ROAS and CPA, retargeting.
Write or iterate ad copy at scale: headlines, descriptions, primary text, variations for creative testing. Any platform.
Plan or improve a static gated asset traded for an email: checklist, template, ebook, content upgrade.
The measurable half: what the page does, what the signup does, and how you know which change caused what.
Improve conversion on a marketing page or form: home, landing, pricing, lead capture, or a bare URL to review.
Run a research-led conversion audit and design evidence-based A/B tests: funnel mapping, objection handling, experiment hypotheses.
Improve post-signup activation: first-run experience, empty states, setup checklists, aha moment, time to value.
Improve completion of a signup, registration, or free-trial flow: fields, steps, friction, abandonment.
Design or improve in-product upgrade moments: paywall, upgrade modal, feature gate, usage limit, trial expiry.
Design or improve interrupt elements: popups, modals, overlays, slide-ins, sticky bars, exit intent, triggers.
Design and run an A/B or split test, or build an experimentation program: hypothesis, sample size, significance, backlog.
Set up or audit tracking and measurement: events, conversions, UTMs, tag manager, attribution, tracking plans, or events that are not firing.
Choose or audit startup metrics: One Metric That Matters, vanity metrics, benchmarks by stage.
Why people stay, why they leave, and what a product has to do in the first week to earn the second.
Fix drop-off with behavior design (B=MAP): motivation vs ability vs prompt, designing the smallest next step.
Reduce churn and keep subscribers: cancel flows, save offers, exit surveys, pause, dunning, failed-payment recovery.
Build habit-forming loops: trigger, action, variable reward, investment. Notifications, streaks, why users stop returning.
Launch and grow a product that is worthless without other users: marketplaces, social, collaboration. Atomic networks, the hard side, liquidity, critical mass.
Design for intrinsic motivation with autonomy, mastery, and purpose: team incentives, engagement loops, gamification that isn't working.
Building a pipeline that is repeatable, asking questions that get honest answers, and closing without theatre.
Build an outbound B2B sales motion: SDR, AE and CSM roles, cold email prospecting, qualification, pipeline targets.
Prepare or run a negotiation: salary, vendor, contract, pricing dispute. Tactical empathy, calibrated questions, BATNA.
Run customer interviews that avoid false positives: no pitching, ask about past behavior, get commitment. The Mom Test.
Deciding what to build before building it, and finding out early when the answer is nothing.
Scope an MVP, run validated-learning experiments, and decide pivot or persevere with Build-Measure-Learn.
Set up a weekly customer-interview habit tied to outcomes: opportunity solution trees, assumption mapping, choosing which experiment to run.
Set product direction with empowered teams and dual-track discovery: product vision, strategy, roadmaps, deciding what to build next.
Run a 5-day sprint to prototype and test an idea with real users before building it: map, sketch, decide, prototype, test.
Design by hypothesis and experiments, not deliverables: assumption mapping, design studio, outcomes over output.
Shape and ship lean products on fixed-time variable-scope cycles: appetite over estimates, betting, saying no. Shape Up, Rework, 37signals.
What good looks like on a screen, argued from principle rather than taste.
Fix visual hierarchy, spacing, color, depth, and design tokens in a UI. Grayscale-first polish, dark mode.
Evaluate an interface for usability problems: heuristic review, confusing flows, forms, navigation, severity.
Diagnose why an interface confuses people: affordances, signifiers, feedback, mental models, error prevention.
Design microinteractions: button and input feedback, loading states, toggles, validation, state transitions.
Choose, pair, and implement typefaces: type scale, line height, responsive type, web font loading.
Build an award-level immersive web experience: cinematic scroll, motion, dramatic type, Awwwards-grade craft.
Design native iOS and iPadOS interfaces to Apple's HIG: navigation, tab bars, sheets, SF Symbols, safe areas, dark mode, accessibility. SwiftUI and UIKit.
Critique a product, feature, or roadmap for focus and ruthless simplicity, and decide what to cut.
Strategy that is a decision rather than a wish, and the operating habits that make it happen.
Write or audit real strategy: honest diagnosis, guiding policy, coherent action. Spots goal lists and fluff posing as strategy.
Run a company on EOS (Traction): V/TO vision, accountability chart, quarterly rocks, scorecard data, Level 10 meetings, IDS issue solving.
Manage for team output: leverage, one-on-ones, meeting cadence, delegation, OKRs, performance reviews, task-relevant maturity.
Design team boundaries and ownership for fast flow: stream-aligned and platform teams, cognitive load, Conway's law.
For the employees that touch code: how to write it, structure it, and keep it running.
Review or refactor code for readability: naming, function size, code smells, comments, error handling, unit test quality.
Decide where code belongs and which way dependencies point: layers, boundaries, ports and adapters, business rules off the framework.
Model software around the business domain: bounded contexts, aggregates, ubiquitous language, domain events, service boundaries.
Judge whether a module, class, or API pulls its weight: deep vs shallow interfaces, information hiding, complexity red flags.
Engineering judgement: DRY, orthogonality, tracer bullets, reversibility, estimation, build vs buy.
Restructure existing code without changing behavior: name the code smell, pick the named transformation, apply in test-backed steps.
Change or test code that has no tests: seams, characterization tests, sprout and wrap, breaking dependencies.
Design a scalable distributed system: load balancers, caching, queues, database scaling, capacity estimates, monolith vs microservices.
Choose and design data systems: SQL vs NoSQL, replication, partitioning, transactions, consistency trade-offs, stream processing.
Make a system survive production: circuit breakers, timeouts, retries, bulkheads, health checks, cascading failure, safe deploys.
Speed up page loads at the network and rendering layer: Core Web Vitals, HTTP/2, caching, resource hints, critical rendering path, WebSocket vs SSE.
1,000 credits, no card. Ask an employee to score what you sell today and tell you what it would change.
Or see the playbooks for the same thing done step by step.