Six Thinking Hats · IN SHORT
Six Thinking Hats separates facts, feelings, risks, benefits and possibilities. The skill works through each perspective before bringing them together in a reasoned next step.
Your brief
Fictional colleagues know the rent but not the notice period. One is excited and uncertain about travel time.
Possible draft
The rent belongs with facts; the missing term remains an open question. Excitement is separate from reasoned benefits such as easier discussion. Alternatives might include a day pass or a trial arrangement, subject to actual availability.
01 / THE FOUNDATION
You receive six distinct perspectives, up to five risks and benefits, at least three alternatives and a concluding synthesis.
02 / SEE HOW IT WORKS
Fictional colleagues know the rent but not the notice period. One is excited and uncertain about travel time.
The rent belongs with facts; the missing term remains an open question. Excitement is separate from reasoned benefits such as easier discussion. Alternatives might include a day pass or a trial arrangement, subject to actual availability.
The skill must not invent facts or other people’s feelings. Ask the people involved rather than attributing a reaction to them.
03 / GET STARTED
The blue hat opens; the white hat separates supplied facts from missing information.
Give feelings, risks, reasoned benefits and creative alternatives their own sections.
The blue hat connects the findings and identifies what still needs investigation.
04 / MAKE IT YOURS
The English ZIP contains the SKILL.md instructions and is free to download without an account. The file declares the MIT licence; retain its source and licence notices when adapting or sharing it. Need help applying this skill? Book a call.
Full informationDownload. Apply. Refine.
File language: English
Explore the details at your own pace.
You put a decision, a plan or a dilemma in front of Claude, and it puts six thinking hats on. One at a time, never mixed. White for the facts and the gaps in the facts, red for the feeling with no justification, black for the risks, yellow for the benefits with reasoning under them, green for the alternatives and blue for the process and the synthesis. What you get back is not an opinion with arguments around it, but six separate blocks plus a conclusion that is allowed to say the decision is not ready yet.
the lines above come from the zip on this page · SKILL.md is 6,188 bytes
This is the worked example that sits literally in the SKILL.md, shortened here. A business owner cannot decide about a second location. Watch the white hat in particular, because it holds not only what is known but also what is missing. And watch the ending: the skill deliberately takes no side here.
The AI Six Thinking Hats Analyser is a free skill from the TheSEO skill library. A skill is an instruction file, SKILL.md, that gives an AI assistant a fixed way of working for one task. Not software you install, not a subscription, not a connection to your systems: a text file that tells Claude which six thinking hats to put on, in which order, how many points may go under each hat at most or at least, and what it may under no circumstances do. You download the zip at the top of this page, put it in your Claude environment, and from that moment Claude analyses a question along this framework instead of giving an opinion about it.
The problem the skill solves is described in the file itself, and it is recognisable. Normal thinking runs together. Facts, feelings, criticism and opportunities collide all at once, and that makes thinking slow and emotional. You have lived through it in every meeting that stalled: somebody names an opportunity, somebody else shoots it down with a risk, a third brings in a feeling nobody wants to admit out loud, and after an hour nobody has changed their mind. Not because the subject was too hard, but because five kinds of thinking were in the same room at once.
Edward de Bono's solution is as simple as it is counterintuitive: let everybody look at the same aspect at the same moment. That is called parallel thinking, and the file says briefly and clearly that it gives depth instead of argument. For an AI that principle works just as well as for a team, because an AI allowed to say everything at once produces a smooth paragraph in which the risks cancel out against the opportunities.
This analyser is meant for anybody stuck in a choice where facts, emotions, risks, opportunities and creative alternatives are running into each other. That is how it is put in the frontmatter, along with preparing a meeting, strategy sessions, conflict situations and brainstorms where the ideas are not being structured. It is part of the skill library we make available free from our AI and automation service: 100 skills, with no account and no sales email afterwards.
The core principle gets its own heading in the SKILL.md, and not for nothing. The whole method stands or falls with the idea that you can split thinking into modes that do not get in each other's way. Four consequences of that show up in the way the skill is built.
Criticism finally gets a place. In an ordinary conversation criticism is socially expensive: whoever names a risk looks like they are against the plan. Under the black hat criticism is the assignment, and that takes the charge out of it. A hard limit is attached to it straight away: not criticism for its own sake, but concrete risk analysis. Five risks with a reason why, not fifty objections.
Feeling does not have to be justified. The red hat is the one people find oddest and the one that delivers most in practice. On that point the instruction is explicit: no justification needed, it may be irrational, and contradictory feelings are allowed. The example holds a line no analysis would have produced: is this ambition or escaping the first business. That kind of unspoken resistance steers decisions, whether you write it down or not.
What you do not know becomes visible. The white hat asks not only for facts but also for the gaps. In the example the unknown rent, the unknown competition, the unknown availability of staff and the unknown ongoing return sit neatly in a row. That list is often the real result of a session, because as long as those four boxes are empty every conclusion is a guess with formatting around it.
There is a hat for the process itself. The blue hat does something the other five cannot: it looks at the analysis instead of at the subject. It restates the question at the start and decides at the end what the next step is, including pointing out the hat that should stay on longer. Without that meta hat the method produces six interesting lists and no decision.
In diff form, with lines from the example in the skill: what an ordinary analysis makes of it, and what is left under the hats.
Every hat has its own job and its own limit. Those limits are the most interesting part of the file, because that is where it says what does not belong under a hat. The column on the right sums up per hat what is and is not allowed, as it stands in the SKILL.md. And note: the hats deliberately get no six colours here, because the file forbids colours in the output.
A skill is only as good as its instructions, so we simply describe them here, without making what is there look better. The file opens with frontmatter: the name ai-six-thinking-hats-analyser, version 1.0.0 and licence MIT. The description under it determines when Claude picks the skill up by itself: on six thinking hats, six hats, de Bono and parallel thinking, but also on lines such as give me all the perspectives, help me think, analyse this from different angles, show me all sides, structure this decision and simply I cannot decide.
Then comes the role. Claude is a structured thinker applying Edward de Bono's framework from 1985, and the assignment is to look at a question from six separate thinking hats. There is a sentence with it that sums up the whole file: they go on one at a time, never mixed. After that every hat gets its own block, exactly as in FIG.02 above.
The method counts seven steps, in a fixed order. Step 1 is restating the question, which is where the blue hat starts: summarise the core of it, in one sentence. Step 2 is the white hat, with a list of known facts and a list of missing facts, and no interpretations. Step 3 is the red hat, with first feeling, gut reaction and intuition in short sentences, where contradictory feelings are allowed too.
Step 4 is the black hat, with at most five concrete risks, each with a short explanation of why. Step 5 is the yellow hat, with at most five concrete benefits, each backed up with why it works. Step 6 is the green hat, with at least three alternatives or creative variations on the original approach. Step 7 is the blue hat synthesis, with a recommendation or clear advice on the decision, plus a note on which hat you should keep on longer if uncertainty remains.
The output structure is fixed as well. Always clear headers per hat, no emoji hats but the name of the hat as a heading, and close with a blue hat synthesis paragraph. That last requirement looks cosmetic but is not: it forces a paragraph somewhere in which the separate blocks come together.
Finally five style rules. Short and sharp, no walls of text. Every hat gets its own block, never merged. English, unless the input is in another language. Fill in no assumptions, work only with what the user brings. And: if a hat cannot be applied for lack of information, say so under the white hat as missing data. Those last two rules together make the skill usable, because they stop you getting back a neatly filled in analysis that is largely invented.
The instruction closes with a source note of two titles and a separate block headed Do Not, holding four things the skill never does under any circumstances. Those four are worked out in FIG.04 further down.
The order is not a suggestion but the method itself. Blue opens, because without a sharp question you analyse the wrong thing. White goes before red, so the feeling is not taken for a fact. Black before yellow, so the optimism does not go untested. Green after the criticism, because alternatives only become useful once you know what is wrong. And blue closes.
The source note in the SKILL.md is short: two entries. That suits a framework with a clear inventor, and it makes checking easy.
De Bono, 1985. Six Thinking Hats, published by Little, Brown and Company. This is the source of the method itself, and the reason the hats are six and not four or eight. De Bono was a physician and psychologist and also coined the term lateral thinking, which comes back in the file as the description of the green hat. His core idea is that thinking is a skill you can structure rather than a talent you either have or do not have, and the hats are the practical tool for it.
The Decision Lab. The instruction points to our own reference guide on Six Thinking Hats. That is a reference source on behavioural science and decision making, and the reason it is there is useful to know: the framework is treated there as an instrument for deciding, not as a creativity trick. That explains why the skill ends with advice on the decision and not with a list of insights.
Even without installing the skill you can get somewhere with this principle: take your own dilemma, write six headings under each other, and force yourself to write down under the white hat what you do not know as well. If the dilemma turns out to be where to compete rather than what to decide, the Blue Ocean Strategy Coach takes that question on, and if the white hat keeps returning that you do not actually understand the subject well enough, the AI Feynman Technique Coach is the one that tests it.
If you then want to test the thinking harder, the Red Team Analyst skill takes the role of the black hat and turns it into full opposition, and the Socratic Method Coach skill keeps digging into the assumptions that surfaced under the white and red hats.
The SKILL.md closes with a block headed Do Not in those words, with four rules in it. Together with the style rules those form the limits of the skill, and those limits matter at least as much as the six hats themselves. In a conversation they play out like this: on the left the request you might make, on the right what happens according to the file.
The zip contains one folder, ai-six-thinking-hats-analyser, with the SKILL.md in it. Installing is a matter of putting that file in the right place, and that place differs per environment. SKILL.md has been an open standard since December 2025, so the same skill also works in Codex, Cursor and Gemini CLI. You are not downloading a Claude file but a working instruction any modern AI assistant can read.
~/.claude/skills/ (or .claude/skills/ in your project)./ai-six-thinking-hats-analyser.AGENTS.md in your repo.AGENTS.md.After that, using it is simple: describe your question, in one sentence if that is all you have, and let the hats do their work. The more facts you give, the fuller the white hat and the less ends up under missing data. Anyone who is new to files like this should first read what Claude skills are and then the step by step explanation of installing Claude skills. If you want to adapt the hats to your own way of deciding, for example with a fixed seventh step for your team, writing a SKILL.md explains how to do that without wrecking the structure. For the wider context on working with AI there is the knowledge base.
This method is at its strongest on a choice you have been circling for a while. A second location, a first employee, a price rise, entering a partnership or ending one. Those questions have all five ingredients running into each other: facts that are partly missing, feelings you do not say out loud, risks you make too big and too small by turns, opportunities that mainly sound attractive, and alternatives nobody thinks of because the conversation is about yes or no. It also works well as preparation for a meeting: take the six blocks in as the agenda, and use the meeting to fill each hat instead of to argue.
There are also situations where you are better off leaving it alone. The first is a question that is factual. Am I legally obliged to sign this contract is not a six hats question, that is a question for somebody who knows the subject. The second is generating something new: if you have no dilemma but a product that needs renewing, the SCAMPER Innovation Coach skill makes more sense, because it generates variants instead of weighing perspectives. For the wider design process around that there is the Design Thinking Coach skill.
The third limit is the most important and sits in the file itself: the skill only works with what you bring. It looks nothing up, it checks no figures and it does not know your market. Give it little information and you get a short white hat with a long list of missing data, and that is exactly what should happen. Reading that as a shortcoming of the skill is confusing a mirror with a lamp.
One more honest limit about the outcome. A blue hat synthesis that says the decision is not ready feels like no answer. It is the most usable answer an analysis can give, because it tells you exactly which two figures to fetch before the rest matters. So do not expect a knot to be cut for you: expect an ordered picture plus the next step.
This skill is the free do it yourself version of work we also deliver as a service. It stays complete and without any catches, but understand well what a skill is: it teaches your AI how to do something, while every new session starts empty. A skill is not the engine and not the memory. You prompt, you supply the facts again every time, you decide. If you want that differently, there are two next steps: hand over the engine, or sort out the memory.
Step 3 deserves more than a paragraph, because this is the difference between a clever chat and a system you can build on. With a decision skill that difference stands out even more: the white hat asks for facts, and without memory you type them in again every time. Worse, you forget which decision you took last month and why. Jarvis is the organisation brain: it remembers what your AIs need to know, divides the work and keeps track of what happened.
We have been running on this system ourselves for months. Every agent session, every task and every decision is logged in it and can be read back. A new session therefore does not start blank: it first fetches the recorded decisions, the running projects and the latest changes, and carries on where the previous one stopped. So we are not describing a promise but the way we work every day.
Amounts are in euros and include Dutch VAT. All plans are listed at jarvis/prijzen (in Dutch). You choose your own plan. Paying online opens once the last checks are finished; until then we set your account up on request. For business arrangements and bespoke work there is a conversation.
Six hats weigh what is on the table. For inventing new options and for really breaking a plan, other frameworks make more sense, and those sit as separate skills in the same library. These are the skills that border directly on this one according to the relationship map, each with the reason attached.
Four skills that take one hat as a starting point instead of as one of six.
SCAMPER Innovation CoachSeven lenses that generate variants on something you already have, where this skill weighs existing options.SKILL Design Thinking CoachThe wider design process: understand the user first, weigh solutions afterwards.SKILL Disney Method CoachThe same separation of thinking modes, but in three rounds: dreamer, realist, critic.SKILL Feynman Technique CoachFor the white hat: checks whether you can explain in plain words what you claim to know.SKILLFive risks are a start. To really break a plan open, these are the skills you want.
Red Team AnalystOrganised opposition: attacks your plan as if it belonged to a competitor.SKILL Pre-mortem AnalystSuppose it has failed a year from now: what caused that. A good follow up to the black hat.SKILL Socratic Method CoachKeeps digging into the assumptions that surfaced under the white and red hats.SKILLAnd the blue hat points at the next step. These skills are about cutting the knot.
Cynefin DeciderWorks out first what kind of problem you have, and therefore which approach fits.SKILL WRAP DeciderFour steps against the classic decision errors, useful when it stays stuck on yes or no.SKILL Inverse ThinkerTurns the question around: not how it succeeds, but how it fails for certain.SKILLWhere this skill comes from and what else there is.
The whole skill library100 free skills in one list, sorted by subject.HUB AI and automationThe service behind it: from separate skills to automation that works in your business.SRV AI trainingIf your team wants to learn to run this kind of session themselves.SRV Knowledge baseArticles about SEO, AI and online visibility, searchable.DOCNothing. The frontmatter of the file says license: MIT, and you do not have to create an account or leave an email address. You download a 3.0 KB zip containing one folder and a single file, SKILL.md, and that is the complete skill. There is no paid version and no sales email follows.
Yes. SKILL.md has been an open standard since December 2025, so the same file works in Codex, Cursor, Gemini CLI and other tools that follow the standard. In Codex you unzip it into .agents/skills/ in your project, or into ~/.agents/skills/ for all your projects; Codex has supported SKILL.md directly since the open standard of December 2025. Pasting the contents of SKILL.md into your AGENTS.md still works too. The instructions are plain readable text, so any assistant that accepts instruction files can handle it.
The method in the file counts seven steps and the order is fixed. The blue hat starts by restating the question in one sentence. Then come white, red, black, yellow and green, each in its own block. The blue hat closes with the synthesis. So blue goes on twice: once to sharpen the question and once to weigh what came out.
Then the skill fills in nothing. One of the style rules is that it fills in no assumptions and works only with what you bring. If a hat cannot be applied for lack of information, that belongs under the white hat as missing data. The Do Not block says it once more: do not fill in every hat when data is missing, say which information is needed instead.
The file sets three limits. The black hat gives at most five concrete risks, each with a short explanation of why. The yellow hat gives at most five concrete benefits, each backed up with why it works. The green hat gives at least three alternatives or creative variations. For the white, red and blue hats no number is set, but an assignment is: facts and gaps, short sentences, and a synthesis.
You get both, but in a fixed order. The file forbids a personal recommendation without a blue hat synthesis, so the advice only comes after the other five hats have been. If uncertainty remains, the blue hat points out which hat you should keep on longer. In the example in the skill the outcome is that the decision is not ready yet.
When there is nothing to weigh. Six Thinking Hats works on a question where facts, feeling, risks and opportunities are running into each other. For generating new variants on an existing product SCAMPER makes more sense, and for a purely factual question putting six hats on is laborious. The skill also researches nothing: it works only with the information you supply.
A sharp weighing up is pleasant, but most decisions business owners bring in eventually come down to the same thing: too little is coming in to be able to choose calmly. Then the real work sits before the six hats, namely being found by the people looking for you, in Google and in AI answers. The free SEO scan shows in seconds where your site stands. And if you want to talk about what AI can do for your business beyond this, from separate skills to full automation, we simply do that in a conversation.
