Find a workflow, add your context, and run it in ChatGPT, Claude, or Claude Code. Use a skill once, or install it permanently as agent behavior.
Turn dates, budget, bags, timing, and preferences into a flight shortlist worth booking.
SFO to JFK next Tuesday, carry-on only, land before 6pm.
Ranked flights with fee traps, timing risk, and booking handoff.
Pick one useful task, add your context, and run it in ChatGPT, Claude, or Claude Code. No prompt expertise needed. Everyday skills are often the easiest first win because the output is personal and easy to judge.
These are the easiest first wins because you already know what a good answer should feel like.
A realistic weekend itinerary with food, timing, geography, and backups.
Planning with constraints
A flight shortlist that catches fees, timing risk, and booking tradeoffs.
Decision support
A practical read on whether a message looks like phishing.
Safety checks
Three weekly priorities, one tradeoff, and a Friday success check.
Prioritization
A reservation plan with ranked options, timing windows, and backups.
Preference handling
Same loop, higher leverage: add context, run the workflow, and judge the output.
A short cold email with a real reason to reply.
Drafting with context
A one-page brief from a long document.
Extracting signal
A scoped build plan from a rough app idea.
Turning ideas into plans
A long thread reduced to decisions, open questions, and action items.
Structuring mess
A side-by-side comparison of two articles or sources.
Comparing arguments
Many people start with trips, restaurants, planning, or safety because the result is intuitive. Once the loop clicks, work skills use the same pattern: context in, structured output out, human approval before anything important.
A catalog of 332 skills is useful only if people can find the first one. These paths route new users to a small, practical shelf instead of asking them to browse the whole catalog cold.
Start with weekly planning, calendar audits, thread summaries, and reusable personal ops.
Find accounts, write outreach, prep calls, and turn messy deal context into next steps.
Use the lifestyle shelf for flights, trips, restaurants, hotels, and contingency plans.
Compare sources, map markets, pull key numbers, brief meetings, and fact-check claims.
Triage suspicious emails, AI-generated scams, privacy traps, and risky requests.
Scope app ideas, debug logs, review PRs, write specs, and delegate code work safely.
Bundles package the catalog around jobs people already understand: founder work, sales and BD, travel, safety, job search, and assistant operations.
Research markets, find accounts, write updates, pressure-test competitors, and turn messy company work into reusable AI workflows.
Find targets, write outreach, prep calls, map stakeholders, handle follow-up, and audit pipeline quality.
Turn app ideas, bugs, PRs, specs, screenshots, APIs, schemas, and launch plans into safer build workflows.
Flights, restaurants, weekend trips, hotels, group dinners, points decisions, and calm backup plans.
Check suspicious messages, phishing, AI scams, account recovery, contracts, tool stacks, and incident plans.
Resume bullets, interview prep, target companies, negotiation, LinkedIn cleanup, and weekly search tracking.
Start with the flagship target account workflow: a complete BD research sprint that turns a company list into ranked accounts, fit logic, owner hypotheses, next moves, and outreach angles. Batch 003 adds public-market and financial-analysis workflows for earnings, filings, valuation sanity checks, stock theses, ETFs, and revenue quality. Founding Pro includes 286 complete Pro workflows in the delivery library, public catalog previews before purchase, and future Pro batches included for founding customers.
A ranked sponsor table with fit rationale, likely owner, buying signal, first angle, confidence, exclusions, and next enrichment step.
A buyer-ready comparison with claim audit, pricing clarity, integration risk, buying committee checklist, demo script, and trial plan.
A compliance-prep map with likely rule areas, triggers, low-regret next steps, official-source checks, and a counsel handoff template.
An evidence pack with original-vs-derivative source ranking, claim support, caveats, confidence, and next verification checks.
A decision-ready memo with executive summary, evidence, caveats, options, recommendation, owner, and next steps.
A battlecard-style teardown: positioning, proof, weak claims, pricing notes, target customer, and ways to sell against them without sounding petty.
A practical account plan with stakeholder map, likely pains, proof angle, entry points, risks, next moves, and 30-day path.
A buyer-ready proposal with decision context, scoped offer, proof, risks, next steps, and a tone pass against the user's own examples.
A risk review showing data exposure, access gaps, policy misses, redundant tools, quick fixes, and escalation questions.
A skill is not magic words. It's a reusable operating procedure for an AI assistant. The file teaches the model how to do one specific job the way a careful person would do it: what to ask, what to ignore, what to produce, and when to keep a human in the loop. That works for everyday tasks like trips and safety checks, and for work tasks like research, sales, writing, and software planning.
The blanks tell the assistant what details matter before it starts.
The rules tell the assistant what to ignore, verify, or hand back to you.
The format turns a loose chat into a useful plan, brief, draft, or decision aid.
Skillbook teaches by showing its work. Free skills reveal the full file. Pro skills show enough of the structure to understand the method before you unlock the library. Every full skill can run once in ChatGPT or Claude, or become reusable behavior in Claude Code, Claude Projects, or a Custom GPT.
--- name: Draft a cold email for: Founders, BD --- # Role You write cold emails that sound like a person, not like a template. Specific, short, polite, useful. # Steps 1. Read what I tell you about them. 2. Find one true thing to lead with. 3. Make the ask in one sentence. 4. Sign off without flourish. # Rules - Never say "I hope this finds you well." - Never list more than one ask. - Under 90 words. - If you don't have a real reason to write, say so and stop.
332 skills, organized the way real work is organized. Start anywhere. Most people start with whichever one they were going to do today anyway.
Turn dates, budget, bags, timing, and preferences into a flight shortlist worth booking.
Convert an occasion, neighborhood, party size, and vibe into a ranked reservation plan.
A realistic two-day itinerary with food, timing, neighborhoods, and one backup plan.
A short, specific, non-cringey email based on what you actually know about them.
Convert a rough idea into scope, screens, data, milestones, risks, and a first commit plan.
Three priorities, one tradeoff, and the meeting you should cancel.
Free is the intentional starter track: everyday and work skills that are useful on their own, and good enough to learn the loop before Pro. Pro includes the founding library: 286 complete Pro workflows in the delivery library, public catalog previews before purchase, and future Pro batches included for founding customers.
An intentional beginner track with everyday and work skills that are useful on their own.
286 complete Pro workflows in the delivery library, public catalog previews before purchase, and future Pro batches included for founding customers.
A done-with-you setup to adapt the library to real work and leave with a usable operating system.
No auth, no dashboard, no nonsense. Read the principles →
Get updates as the Pro workflow library fills in. This also tells us which bundles and work products to polish first.
A prompt is usually a wish. A skill is a reusable operating procedure: inputs, process, boundaries, examples, and the shape of the output.
Pick the job you were already going to do, fill in the blanks, then run it in chat or install it as reusable behavior.
The starter library is deliberately beginner-friendly: everyday tasks, work tasks, and enough structure to learn the loop without paying first.
Showing the raw .skill.md body matters because you can inspect it, edit it, save it, and understand what the assistant is being asked to do.
The best agent workflows are not reckless automation. They are scoped, reversible, and clear about when you should approve the next step.
That's the whole pitch. If a free skill does not pay for itself in twenty minutes of saved work, the rest are not worth your money either.