How to Use This Vault
Each prompt below assigns the AI a specific coaching role, feeds it a realistic scenario in brackets, and demands an output you can act on the same day. The prompts are sequenced to match a real sprint arc: audit first, then planning, then stories, then ceremonies, then metrics, then the defense rehearsal. You do not need all twenty-five, the first four alone will change your next sprint.
Two habits multiply every result. First, add one sentence of your own context to every run: the model does not know your team, your deadline, or the stakeholder who escalated last Tuesday, that sentence is where generic drafts become your coaching. Second, keep the outputs. A retro experiment card the model drafted is only valuable if someone reads it at the next retro opening; a canvas the model pressured into honesty is only valuable if it survives until Friday.
Why Prompt Skills Are Table Stakes in 2026
The adoption floor keeps rising. DORA's 2024 State of DevOps report found 76% of technology professionals relying on AI for at least one professional duty, and McKinsey's global survey put regular generative AI use at 65% of organizations, delivery and engineering among the leaders. When the market is this saturated, having access to AI is not an advantage; knowing what to ask it for is.
The measured gains back that shift. GitHub's controlled study found developers completed a standard task 55% faster with an AI pair, and the effect concentrated on patterned work, exactly the work most agile prompts target. A prompt library is how a team converts that tool-level speed into team-level practice: the same audit, the same canvas, the same rehearsal, run every sprint without depending on anyone's mood or memory.
Sources: DORA State of DevOps Report 2024 · McKinsey Global Survey on AI 2024 · GitHub Copilot research study 2022
Sprint Audit and Baseline Prompts (1-5)
These five prompts produce the before-photo: an honest account of how your process actually runs, which value pairs it serves, and what it quietly betrays. Run them once, save the output, and re-run the audit after a month of real changes, the distance between the two is your progress report. Teams that skip the baseline always regret it at review time.
- 1. Theatre Auditor, score your last sprint 1-10 on output worship, ritual compliance, and tool religion; name the worst tell
- 2. Values Court, put one ceremony on trial against the four Agile Manifesto value pairs and rewrite its format
- 3. Decision Counter, classify every ceremony output from last sprint as Decision, Information, or Ritual; total the hours
- 4. Ceremony Timebox Auditor, find the ceremonies that earn their calendar slot and the ones to shrink or merge
- 5. Baseline Memo Builder, interview you, then write the 150-word before-photo you re-read after 15 days
Planning and Backlog Prompts (6-12)
The planning prompts install the Sprint Clarity Canvas and the backlog funnel: a goal sentence a stakeholder recognizes, honest capacity math, WSJF scoring that out-argues the loudest voice, and stories sliced thin enough to finish. The story prompts then make each slice testable, Gherkin criteria a tester can run without asking a single question. The full user-story workflow with AI shows prompts 11 and 12 inside the complete loop.
- 6. Canvas Builder, force the four quadrants: Goal, Capacity, Risks, Demo Plan
- 7. Goal Sharpener, rewrite weak sprint goals as one-sentence, demo-able outcomes, scored for clarity
- 8. Capacity Reality Check, compute gross versus honest capacity and the focus factor as a percentage
- 9. WSJF Scorecard Builder, score candidates on value, decay, and size; rank the queue
- 10. Decay Detector, separate true time criticality from requester impatience, one sentence each
- 11. Gherkin Writer, turn vague criteria into Given/When/Then plus the four edge cases you missed
- 12. SPIDR Slicer, cut your scariest epic into five vertical slices, blade by blade
Ceremony and Retro Prompts (13-19)
These prompts fix the meetings themselves: the board-walk script for tomorrow morning, the blocker sentences that make escalation work, the review script with a forced decision, and the Rocket retro that lands an Experiment Card with an owner and a success signal instead of a feelings dump. Facilitation craft still matters, the prompts carry the structure so you can carry the room. The retrospective method behind prompt 17 explains the three phases in full.
- 13. Board Walk Scripter, the right-to-left run of show from your real board state
- 14. Blocker Sentence Coach, upgrade I-am-a-bit-stuck into what, since when, tried what, next rung
- 15. Demo-Ladder Builder, script the live demo, the two real scenarios, and four harvest questions
- 16. Harvester Grid Scribe, route raw review feedback into Like, Wonder, Concern, Idea, with destinations
- 17. Rocket Facilitator, the three-phase retro script with your context baked in
- 18. Experiment Card Writer, convert the eternal gripe into hypothesis, change, owner, date, signal
- 19. Five-Whys Digger, stay on the system, stop at a process cause you can test
Metrics, Scaling, and Defense Prompts (20-25)
The final prompts make the numbers honest and the room safe. A cycle-time baseline from your real completion history, a WIP experiment designed with objection scripts, the Now-Next-Later roadmap draft, and, the one that saves the most careers, the objection rehearsal, where the model fires the eight hardest stakeholder questions at your system before the room does.
- 20. Cycle Time Baseline, average, spread, and outliers from your last twenty completed items
- 21. WIP Experiment Designer, the limit, the finish-first rule, and answers to the three objections you will hear
- 22. Queue Hunter, find the single most expensive waiting state on your board
- 23. Roadmap Builder, draft Now-Next-Later with confidence sentences and the footer deal
- 24. OKR Ladder Drafter, link company KRs to team OKRs to sprint goals; flag the drift
- 25. Objection Rehearser, your numbers against the eight hardest questions, answers in one breath
What the Measured Gains Look Like
The evidence on AI-assisted work is unusually consistent for a young field. A large-scale field study published through NBER found generative AI assistance raised support-agent productivity by 14% on average, with the largest gains, 34%, going to the least experienced staff. Nielsen Norman Group's across-task research measured throughput improvements of about 66% for business professionals using generative AI on routine work.
Read the pattern, not just the numbers: AI helps most where work is patterned and least where judgment is scarce. Every prompt in this vault is aimed at the patterned layer, scoring, drafting, structuring, rehearsing, while the conversation, the trust, and the decision stay with the team. Teams that inverse that split, delegating judgment to the model, get the speed without the change.
Sources: NBER field study (Brynjolfsson et al.) 2023 · Nielsen Norman Group research 2023
Safety, Policy, and What Not to Paste
Every prompt in the vault runs on structure rather than records, which is a deliberate design choice and not only a legal one. Structured context, durations, counts, column names, the shape of a problem, produces sharper coaching than raw tickets, while raw customer data creates risk that no output justifies. Check your organization's AI policy before the first paste, then hold the line these four rules draw.
- No customer names, contract values, or personnel issues in general-purpose chat tools
- Paste summaries and structure, not raw ticket dumps or export files
- Use your organization's approved assistant tier where one exists
- Treat every output as a draft for human review, the veto stays with the team
Making the Prompts Stick: Triggers Over Bookmarks
A prompt library becomes a system the same way a gym membership becomes muscle: through scheduled reps, not enthusiasm. The mechanism that works is the trigger pairing, every prompt attached to a calendar moment that already exists, so using it requires no new discipline. Bookmarked lists without triggers are how prompt collections die quietly in a tools wiki.
The trigger pairings that work
Pair the audit prompt with the first retro of each quarter, the canvas builder with planning day, the Gherkin writer with refinement, and the objection rehearser with the night before any stakeholder review. Teams that pair prompts with triggers keep the habit for months; teams that rely on remembering do not survive the first busy sprint. Start with two pairings and expand only when the first two are automatic.
| Trigger moment | Prompt to run | Output it produces |
|---|---|---|
| First retro of each quarter | Theatre Auditor (prompt 1) | A 1-10 theatre score with the worst tell named |
| Planning day | Canvas Builder (prompt 6) | A four-quadrant Sprint Clarity Canvas |
| Refinement | Gherkin Writer (prompt 11) | Given/When/Then criteria plus the missed edge cases |
| Night before stakeholder review | Objection Rehearser (prompt 25) | One-breath answers to the eight hardest questions |
| First Monday of each month | Prompts 1-5 re-run | A diff of the audit against the original baseline memo |
Source: Pairings compiled from ZLearning Agile & Scrum Mastery prompt-vault usage guidance, 2026
The monthly re-run
Once a month, re-run prompts 1 through 5 and diff the output against the original baseline. The delta is your evidence of change, or your early warning that the theatre has reasserted itself. Either way the model hands you the same gift: an unforgiving account of your own process, produced in minutes, that no dashboard will ever give you.
A 30-Day Plan to Turn Prompts Into Practice
Prompt libraries reward sequencing over volume, so run the month in three deliberate phases. Each phase has one goal, a handful of prompts, and a testable outcome, by Day 30 you should have a baseline, a changed planning rhythm, and a retro that lands experiments instead of grievances.
| Week | Prompts in play | Deliverable | Done test |
|---|---|---|---|
| Week 1 | Prompts 1-5 | A baseline memo signed by the team | Before-photo filed where the team can see it |
| Week 2 | Prompts 6-10 | Canvas, honest capacity math, and a WSJF-ranked backlog | Planning runs on the canvas, not on gut feel |
| Weeks 3-4 | Prompts 13-19, then 20-25 | One landed Experiment Card and a flow baseline | Goal is a sentence, board has a WIP cap, retro opens with kept-adjusted-killed |
Source: Plan compiled from ZLearning Agile & Scrum Mastery course materials, 2026
Week 1: the baseline
Run prompts 1-5 on your most recent sprint and file the outputs where the team can see them. Do not change anything yet, the discipline is measuring before tuning. The single deliverable is the before-photo memo, signed by the team, that every later comparison will be judged against.
Week 2: planning on rails
Introduce prompts 6-10 at your next planning and refinement cycle: the canvas gets built before the meeting, capacity math replaces gut feel, and the backlog gets WSJF-scored ahead of the loudest-opinion contest. Expect friction from one stakeholder, the decay detector prompt exists precisely to give that conversation a number instead of a fight.
Weeks 3-4: ceremonies and the first experiment
Run the Rocket with prompts 13-19, land one Experiment Card with an owner and a date, and close the month with prompts 20-25 to establish your flow baseline and rehearse the stakeholder defense. If the month worked, your next sprint goal is a sentence, your board has a WIP cap, and your retro opens with kept-adjusted-killed. When you want the whole arc, prompts, frameworks, and the missions that force the reps, Agile & Scrum Mastery runs it in fifteen days with a fresh prompt pack each day, and Day 1 is free with any account.
Frequently asked questions
▸Do these agile and scrum prompts work in any AI tool?
Yes, they are written for ChatGPT, Claude, and Gemini, and work in any capable assistant including Copilot. The structure matters more than the tool: each prompt assigns the AI a role, feeds it your real context in brackets, and demands a specific output format. Replace the square-bracket placeholders with your details and run.
▸What is the best prompt to start with?
The Theatre Auditor. Paste how your last sprint actually ran, ceremonies, attendance, outputs, and ask for a 1-10 score on output worship, ritual compliance, and tool religion. Every other improvement builds on knowing where your process performs agility instead of delivering it, and most stuck teams trace back to skipping exactly this step.
▸Is it safe to paste sprint and backlog data into AI prompts?
Follow your company policy first, and when allowed, paste structure instead of raw records: ceremony durations, point counts, column names, and the shape of the problem rather than ticket-level customer data. Every prompt here is designed to work on structure and summaries, which is both safer and usually produces better coaching.
▸How do I get my team to actually use these prompts?
Start with one prompt in one ceremony, not twenty prompts in every ceremony. The audit prompt in your next retro is enough, let the team watch the model name the theatre they already feel. Adoption follows usefulness, and one honest audit converts more skeptics than a mandate ever will.
▸Can I customize the prompts for my own board or stack?
Yes, and you should, the bracketed context fields exist to be overwritten. Swap in your board's column names, your point scale, and your ceremony timeboxes, and the outputs sharpen immediately. The field guide to AI tools for agile teams maps which tool owns which job, so each prompt runs where it fits best.
▸Do prompts replace scrum training?
No, prompts accelerate the practice, but they cannot teach the judgment behind the frameworks they invoke. A prompt can score your sprint for theatre; understanding why the score matters is what a structured program builds. The fastest path is both: a free agile and scrum course day for the foundations, then these prompts as the daily reps.
▸How do I learn the full system these prompts belong to?
Agile & Scrum Mastery ships a fresh prompt pack with every one of its fifteen days, each inside a mission that forces you to run it on your real sprint, from the Sprint Clarity Canvas through flow metrics to the stakeholder defense. Day 1 is free with any account, and the audit you run on it is yours to keep either way.