13 Claude Skills for Design Sprints That Land

What these skills are

A whole sprint engagement, as installable Claude skills. Thirteen skills for the whole engagement, not just the five days. The arc runs from working out whether a sprint is even the right tool, through framing the challenge, casting the Decider and the seven seats, and booking five real strangers for Friday, then the week itself from map and target to sketch, decision, storyboard, prototype and the five interviews, and finally the Monday after — the part every sprint book skips, and where most sprints quietly end.

One line holds through all of it: a sprint buys a decision people will defend on Monday, not a document.

Claude will generate concepts for the wall on sketch day, and they compete on the same heat map as everyone else's — but nothing of its own reaches the vote unless a human puts their name on it. A concept whose only defender is a file will not survive the meeting where it costs something. Ideas are cheap; commitment is what the client is buying. The heat map, the Decider's call and the five conversations stay human.

Mechanically, each skill is one folder with a SKILL.md file. The 13 are grouped into three phases, and each one gives full value on its own: run one, run a phase, or chain the lot. They work from one Claude Project per sprint, so the brief, the map, the storyboard and the grid stay in one findable place and the week leaves a paper trail instead of a memory.

Download all 13 skills. One zip: ready-to-install skill zips, readable SKILL.md files, and the two blank templates. Free, no signup. The same 13 skills are open on GitHub: polar-bear-org/claude-skills.

The 13 skills you get

The set follows the whole engagement, not just the five days: five skills before the week starts, seven inside it, and one for the Monday after.

1 · Before the week starts

1. Sprint Qualifier

Use when: A client just asked you for "a design sprint" and you are not sure they need one

Output: A two-page qualification memo: the five gates with a one-line verdict each, the Decider answer verbatim, the recommendation in one sentence, the alternative to sell if it is a no, and a draft email to the client you edit into your own voice

2. Challenge Framer

Use when: The sprint is happening and you need to know what it is actually about

Output: The one-page sprint brief: slug, long-term goal, sprint questions, challenge statement, customer, target moment, constraints, variant and dates, the Decider, and the "not this week" list. Every other skill in the pack reads it

3. Sprint Cast Builder

Use when: Two to four weeks out, when you are trying to get people to block a week

Output: The seven seats with real names and the knowledge each brings, the Decider's status and the delegation line, the expert list with slots and two questions each, the pre-read text ready to send, the hour-by-hour calendar for your variant, and a risk list of who has not confirmed

4. Test Recruiter

Use when: Ten days out, and the customers who will see the prototype are still hypothetical

Output: The screener ready to send, the sourcing plan with an owner per route, six booked slots with confirmation status, the incentive and how it is paid, the consent text, and the standby plan. The top line is a live count: booked, confirmed, showed

5. Sprint Facilitator Kit

Use when: Before the sprint, then again mid-week when something in the room goes wrong

Output: The room craft rather than the agenda: timeboxing that holds, protecting silent work, the move for each people problem, energy across five days, and running it remote. Optionally a facilitation notes file — process only, never anything about how a named person behaved

2 · The week itself

6. Map and Target Runner

Use when: The morning the sprint starts

Output: The confirmed or amended goal and questions, the map as numbered steps with actors, the How Might We notes clustered with vote counts, the expert quotes worth keeping, the target in one sentence, and the Decider's reason verbatim. Written before anyone goes home

7. Sketch Day Runner

Use when: The night before sketch day, that morning, and again once the human sketches are up

Output: The Lightning Demo list with every source named, the human sketch count with the titles their authors gave them, the machine concepts with the cluster or constraint each came from, and the adoption record: who put their name on what. Never who drew which human sketch

8. Decision Day Runner

Use when: The morning you have to choose between sketches

Output: The heat map's clusters in words, every concern captured as a question, the straw poll result, the supervote with the Decider's verbatim reason for each dot, the rumble or all-in-one call, and the gap between room and Decider if there was one. No record of who put which dot where

9. Storyboard Builder

Use when: Straight after the vote, while the room is still together

Output: Fifteen numbered panels with what is on screen, the real words in it, and what the tester is meant to do; then the map from panel to sprint question, and the gaps the room chose to accept. With the prototype plan, this is the whole build spec

10. Prototype Plan Writer

Use when: The night before build day, or first thing that morning

Output: The panel list marked real, facade or never seen; the tool call; four roles with names; the hour-by-hour build plan with the trial run in it; all the prototype content written out; and the fix list the trial run produces

11. Test Script Writer

Use when: Build day afternoon, once you know what the prototype does

Output: The five-act script with the exact welcome and prototype-introduction wording, your tasks mapped to sprint questions and storyboard panels, the allowed questions on a card the interviewer can hold, the debrief questions, and the note grid set up for your week

12. Friday Debrief Synthesizer

Use when: Friday, five o'clock, before anyone goes home

Output: The grid as a table, patterns confirmed by the room with their counts, single observations kept separate, each sprint question marked yes, no or unclear with its evidence, whether the result changes the bet or the detail, and an honest confidence statement. Testers stay numbered

3 · The Monday after

13. Monday After Planner

Use when: The following Monday, and not one day later

Output: A one-page decision record ready to send, plus a two-week plan: the road you are on, three to five items with named owners and dates, the evidence the sponsor will ask for before funding a build, and the follow-up meeting date. If there is no named owner, the file says so at the top

How a sprint flows through the pack

One Sprint HQ project carries the engagement. sprint-qualifier runs before the proposal goes out, and is built to end in "no" without embarrassment; challenge-framer then writes the sprint brief every other skill reads first. sprint-cast-builder and test-recruiter run in parallel, not in sequence, because five strangers take longer to book than seven colleagues. Inside the week, map-and-target-runner captures day one while you listen, sketch-day-runner scouts the demos and adds its concepts only once the human round is sealed, decision-day-runner holds the mechanics on the clock, and storyboard-builder runs in the same room straight after the vote. prototype-plan-writer and test-script-writer turn that into something a stranger can meet on Friday, and friday-debrief-synthesizer will not let the room go home without an answer.

Then monday-after-planner, which is the skill most sprints are missing. Friday's answer becomes a written decision, one named owner, a two-week plan with dates, and an honest list of the evidence the sponsor still needs before funding a build. Run it on the Monday: every day you wait, the room's willingness leaks away.

Two blank templates ship with the pack and are filled by the skills that own them — sprint-brief.md, written by challenge-framer and read by everything else, and friday-notes-grid.md, which the room fills live during the five interviews and friday-debrief-synthesizer reads at the end of the day.

The pack runs sprints; it does not do discovery. A sprint that needs personas has already failed qualification, and sprint-qualifier names research as the alternative to sell instead — that work lives in the Design Thinking Pack.

Setup guide

1. Download the pack. One zip: an install folder with 13 ready-to-upload skill zips, a skills folder with the same skills as readable files, and the two blank templates.
2. Install your skills. In Claude Code, add the marketplace and install design-sprint-pack, and all 13 load at once. In Claude, turn on code execution in Settings, Capabilities, then go to Customize, Skills and upload one zip per skill from the install folder. Prefer working from files? Add the SKILL.md files to your Project knowledge instead; it works, just less cleanly.
3. Make one Project per sprint. Call it Sprint HQ, and keep one pinned chat per skill, named for the skill, so the sketch day chat does not fill up with recruiting logistics. Every output is saved as [artifact]-[sprint-slug].md, where the slug is set once by challenge-framer and never changed again, and later skills read the earlier files by those names. Then write "run sprint-qualifier" — before you send the proposal, not after.

Where to start

The quality bar

Every skill in the pack holds the same standard, the one we hold when we run a sprint ourselves:

• A sprint buys a decision people will defend on Monday, not a document: the heat map, the Decider's call and the five conversations stay human
• Ideas can come from anywhere, Claude included — but a machine concept that draws heat comes off the wall unless a person puts their name on it, even if it had the most dots
• No concept reaches the wall before the human sketch round is sealed: anchoring is not a preference, it is what will happen
• Claude never votes, never ranks the sketches and never breaks a tie, including on the concepts it generated the day before
• Testers are real strangers: never simulated, never roleplayed, never profiled, numbered in every file and described by what happened on the screen
• Five sessions never become a percentage, a score or a rating: four out of five is four people
• Anonymity holds after the week too — no record of who drew which sketch or placed which dot, and no reconstructing it later
• No invented benchmarks and no manufactured yes: a qualification that ends in "no", with the alternative to sell instead, is the pack working
• Nothing evaluates a named person: a retro looks at what slowed the week down and what qualification missed, never at who underperformed

Who made this

Polar Bear is a people ops consultancy for human-size teams (20 to 200 people). Built by ex-McKinsey founders with a dream to make AI work for People, not instead of them. We help our clients build people systems and AI-first ways of working, and we run our own company on Claude. This pack is the free, self-serve version of how we work.

The pack carries one sprint at a time. When you want your whole team working this way, AI carrying the overhead so people do the thinking, across sprints, workshops, and everyday work, that's what we build with clients.

Meet Pauline. Weighing up a sprint week? Bring the client's ask, the calendar you cannot fill, or the sprint that ended without a decision. Book a 30-minute call · Pauline on LinkedIn