Guides published aug 27, 2026
In this guide, you will learn how to give your AI skills that make every decision easier. You'll use an /unstuck tool, a /decide tool, and more to turn problems into solutions.
You will build a repeatable decision workflow that turns a stuck problem into focused questions, challenged assumptions, and a concrete action. Depending on the skill you choose, that output could be a five-post publishing plan or a project decision with a clear recommendation and revisit date.
/business-brainstorm is a useful extra skill when you have a promising idea but need to shape and pressure-test it. It is not part of the main decision framework, but it can push a loose concept toward a differentiated product:
/business-brainstorm
I want to launch [PRODUCT OR BUSINESS] for [CUSTOMER]. I want to compete in [CATEGORY OR PRICE POSITION] without copying [REFERENCE BRANDS]. Help me develop a differentiated hero product, pressure-test the positioning, and outline likely pricing and cost considerations.We tested this with a premium blanket idea. /business-brainstorm came back with a two-sided product, name ideas like Sunday Weight, and a hook around a blanket that stays in place.
In the video, we followed up with:
Turn the recommendation into a one-page product brief and draft a sourcing RFQ. Use an initial order target of about [QUANTITY] units rather than assuming a large production run.Open Terminal, start Claude Code, and add the makerskills marketplace:
claude
/plugin marketplace add coreyhaines31/makerskillsThen install the library:
/plugin install makerskills@makerskillsIf the marketplace is already installed, Claude Code may simply confirm that nothing needs to change. You can also open the marketplace inside Claude Code, find makerskills, and install it from the menu.
Pro tip: Start Claude Code before pasting the /plugin commands. They are Claude Code slash commands, not shell commands.
Open the repository's routing table before you start. It organizes the library by the job you are trying to do, which is faster than opening every skill one by one.
Two useful starting points are:
In Claude Code, type /, choose the skill, and describe the outcome you need. Start with the skill that matches the decision in front of you instead of trying to test the whole library at once.
Pro tip: Run the skill inside a project or session that already has context on your work. It will ask more specific questions and make more relevant recommendations.
Give the skill the real problem, the result you want, the constraint slowing you down, and any deadline. For example:
/unstuck
I am stuck on [PROJECT OR PROBLEM]. I keep spending time on [ASSUMPTION OR BOTTLENECK] instead of [DESIRED ACTION]. I want to reach [CONCRETE OUTCOME] by [DEADLINE], with these constraints: [CONSTRAINTS]. Help me challenge what I am assuming and choose the lowest-friction next move.Answer the follow-up questions plainly. The value comes from letting the framework test whether the apparent blocker is the real one.
We saw this when social-fetch could retrieve a Hacker News post but could not pull an X post. Its next route required a paid scraping API, which would have turned the tutorial into a costly, gray-area demo.
We used /unstuck to challenge our assumption that we needed the social-fetch demo. It pointed us toward the more useful decision workflow instead.
The same skill helped with a local newsletter that had gone nearly three weeks without consistent social posts because the design never felt finished.
It challenged the value of polish and proposed five low-friction formats: a plain-text post, a newsletter screenshot, a phone photo, a local list, and a neighborhood question. It then recommended batching all five into a scheduler.
Pro tip: If a tool sends you toward paid access, brittle scraping, or a workflow you do not want to own, treat that as new evidence. Revisit the original goal before adding more setup.
Once you understand the pattern, try a different skill on a second problem. /decide works well when both options are plausible but you only have room for one. Give it the options, the deadline, the cost of being wrong, and how reversible the choice is:
/decide
I need to choose between [OPTION A] and [OPTION B] by [DECISION DATE]. I only have capacity for one. Option A offers [UPSIDE] but requires [COST OR RISK]. Option B offers [UPSIDE] but has [TRADEOFF]. Interview me, challenge the assumptions behind the choice, recommend the next move, and give me a date to revisit the decision.We tested /decide with two hypothetical client projects. It asked how much the decision mattered, tested whether it was reversible, recommended the new work stream, and set a future revisit date.
Keep going
Members get all guides, courses, live sessions, and $1,000+ in partner perks.