The Practice Room · runs in your own AI
Your first rep shouldn’t be the real one.
Each month of the curriculum teaches one move. This room is where you run it out loud: copy a prompt into the AI you already use, play the scene against a counterpart who pushes back, and get scored with receipts. This page stores nothing and sends nothing.
Why a practice room
The calendar is a terrible practice schedule.
A charged conversation arrives a few times a month, at full stakes, with no warm-up. Skills set through repetition at rising difficulty with feedback after every attempt, and real life will never supply enough of those reps on its own. So this room borrows the counterpart. You run the month’s move against an AI playing a colleague with a plausible version of events, as many times as you want, at whatever heat you choose. Every session ends the same way: your weakest turn, rewritten in the shape the month teaches. The comparison is where the learning lives.
Pick the scene and the heat
Use a scene seed or write your own, scrubbed of real names and details. Choose a level and the prompt updates itself.
Copy the prompt into your AI
Paste it into whatever AI you already use and run the scene in your own words. Type DONE when the conversation is over.
Read the score, then the rewrite
Four criteria, each backed by a quoted line from your transcript, then your weakest turn redone in the month’s shape. The rewrite is the part worth rereading.
Difficulty
Five levels of heat.
A counterpart who folds the moment you speak teaches nothing. The ladder raises the resistance in steps, so each rep sits slightly above the last one.
The counterpart is cooperative and the stakes are low. You are learning the shape of the move.
They push back once, plausibly, then engage if you handle it well.
They hold a reasonable version of events from their side for most of the scene. You work for every turn.
They are in a threat response: faster, sharper, harder to reach. Your regulation is being tested as much as your words.
A power dynamic, a second conflict type underneath the visible one, or a beat you could not have prepared for. Closest to the real thing.
Week 1, run levels 1 and 2. Week 2, levels 2 and 3. Week 3, levels 3 and 4. Week 4, levels 4 and 5. Once a week, rerun a scene from an earlier month, one level higher than you left it. Log each scene as a rep.
Scoring
The same four criteria, every month.
The weekly reflection asks the same questions every week because the repetition builds the muscle. Scoring works the same way. Each month’s move sharpens one criterion in particular, and every conversation needs all four.
Regulated
You created a gap between trigger and response and chose your moves on purpose instead of automatically.
Fact-grounded
You spoke to observable behavior and owned your interpretations as interpretations.
Understood first
The other person would say you genuinely got their side before you advocated yours.
Landed something real
The conversation ended with a concrete next step, a genuine request or repair, or a clean, honest close.
Every prompt requires your practice partner to quote the exact line of yours that earned each score. A score without a quote is an opinion, and opinions do not teach.
Ground rules
Three rules before you step in.
Change names, drop identifying details, keep the dynamic. The scene needs the shape of the situation, never the confidential parts. This page stores nothing and sends nothing; the prompt runs in whatever AI you already use, under that tool’s terms.
Reps spread across days set deeper than a marathon session. Level up across the month.
Scenes involving safety, discrimination, coercion, or abuse need real support first, and every prompt instructs your practice partner to stop those scenes. The curriculum’s own boundary applies here too.
The rooms
One room per lens.
Each room carries the month’s move, scene seeds, a heat picker, the full prompt, and three retrieval reps. Start in the month you are in.
Month 01 · Practice room
Under the Hood
When activation begins, run STOP: stop, take a breath, observe what is happening inside you, proceed deliberately. The gap between trigger and reaction is the whole skill. Everything you say after the gap counts double.
Recognition cue Your body tightens, your jaw sets, your breath shortens, or your listening shuts down.
Scene seeds
- A colleague copied your manager on an email implying you missed a deadline. The deadline changed, and they knew it. They have just walked up to your desk.
- In a team meeting, a decision lands on something you feel ownership of, made without your input. The person who made it asks what you think.
- A new team member questions the process you built, in front of others, in a tone that suggests they think it is broken.
You are my rehearsal partner for a conflict conversation. We will run a scene, then you will score it. Follow every rule below exactly. THE SCENE [Paste a scene seed here, or describe your own situation with names changed and identifying details removed.] YOU PLAY The person on the other side of that scene. You have a reasonable version of events from your side. Decide privately what it is and stay consistent with it. I AM PRACTICING Running STOP when activation begins (stop, take a breath, observe what is happening inside me, proceed deliberately), then responding from a regulated state instead of an automatic one. DIFFICULTY · LEVEL 2 OF 5 Push back once, plausibly, early in the scene. If I handle it well, engage genuinely. SCENE RULES 1. Open the scene with the triggering line itself. Make it genuinely sting at the difficulty level set above. 2. Stay in character until I type DONE. Speak only as your character, a few sentences per turn, the way a real person talks. 3. Invent nothing dramatic. If I ask about something the scene does not cover, improvise something small and ordinary. 4. Resist plausibly, never cartoonishly. Your pushback comes from your version of events. 5. Keep the scene to roughly six to ten exchanges. If I have not typed DONE by then, ask whether I want to keep going or score it. 6. If the scene moves toward harassment, discrimination, threats, or abuse, stop it, drop character, and say that this belongs with real support rather than rehearsal. WHEN I TYPE DONE Drop character. Score my side of the conversation only, against the four criteria below, 0 to 2 each. Quote the exact line of mine that earned each score. No quote, no score. Where I lost points, quote the line where it happened. 1. REGULATED. I created a gap between trigger and response and chose my moves on purpose instead of automatically. 2. FACT-GROUNDED. I spoke to observable behavior and owned my interpretations as interpretations. 3. UNDERSTOOD FIRST. Your character would say I genuinely got their side before I advocated mine. 4. LANDED SOMETHING REAL. The conversation ended with a concrete next step, a genuine request, or a clean, honest close. Also name which SCARF domain your opening line was aimed at (status, certainty, autonomy, relatedness, or fairness) and say whether my response suggests I noticed. THEN THE COMPARISON Take my weakest turn and rewrite it the way a skilled practitioner would have run it: a visible beat of pause before any content, a named observation instead of a counterattack, and one clear sentence about what happens next, all from a regulated state. Write it in plain language I could actually say out loud. End by naming the one thing to work on in my next rep. One thing only.
Retrieval reps · answer from memory, then check
Three quick checks on the month’s lens. Commit to an answer before you tap.
Mid-meeting, a colleague says your part of the project is the reason the team is behind. Your jaw sets. What is the first move?
Reveal the why
The content can wait ten seconds. STOP puts a gap between the trigger and your response, and everything you say after the gap counts double. The timeline and the clarifying question both come next, chosen deliberately.
A decision you own gets reassigned in front of the team with no explanation. Which SCARF domains most likely just fired?
Reveal the why
Public reassignment hits your sense of relative standing and your control over outcomes at the same time, and the missing explanation adds a certainty hit. Naming the domain is usually more useful than arguing the decision.
A quiet day. No charged moments anywhere. What does the daily rep look like?
Reveal the why
You are training the move to be available before you need it for something that matters. Small reps on quiet days are what make the move findable when your pulse is up.
This is the first month, so the return rep starts in Month 2. From then on, STOP is present in every scene you run: it is criterion 1, every time.
Month 02 · Practice room
The Voice in Your Head
The 30-Second Start. Externalize the story before it runs the conversation: I noticed [observable behavior]. The story I am telling myself is [interpretation]. Is that accurate?
Recognition cue You find yourself certain about what someone meant, why they did something, or what their behavior says about them.
Scene seeds
- A colleague has gone quiet in the last three meetings on a project you lead. You have concluded they have checked out. You ask them for ten minutes.
- Someone keeps rescheduling your one-on-ones. The story you are carrying is that you are low on their priority list. You open the next conversation with it.
You are my rehearsal partner for a conflict conversation. We will run a scene, then you will score it. Follow every rule below exactly. THE SCENE [Paste a scene seed here, or describe your own situation with names changed and identifying details removed. Include the story I have been carrying about this person.] YOU PLAY The person my story is about. You have a reasonable version of events from your side, and it differs from my story in at least one important way. Decide privately what it is and stay consistent with it. I AM PRACTICING The 30-Second Start: I noticed [observable behavior]. The story I am telling myself is [interpretation]. Is that accurate? My job is to keep the observation at the level a camera would have caught and to own the story out loud as mine. DIFFICULTY · LEVEL 2 OF 5 Push back once, plausibly, early in the scene. If I handle it well, engage genuinely. SCENE RULES 1. The moment my "I noticed" contains a judgment dressed as an observation, push back on it in character: that is not what happened, or that is not what I said. Only engage with my story once the observation is something a camera would have caught. 2. Stay in character until I type DONE. Speak only as your character, a few sentences per turn, the way a real person talks. 3. Invent nothing dramatic. If I ask about something the scene does not cover, improvise something small and ordinary. 4. Resist plausibly, never cartoonishly, at the difficulty level set above. 5. Keep the scene to roughly six to ten exchanges. If I have not typed DONE by then, ask whether I want to keep going or score it. 6. If the scene moves toward harassment, discrimination, threats, or abuse, stop it, drop character, and say that this belongs with real support rather than rehearsal. WHEN I TYPE DONE Drop character. Score my side of the conversation only, against the four criteria below, 0 to 2 each. Quote the exact line of mine that earned each score. No quote, no score. Where I lost points, quote the line where it happened. 1. REGULATED. I created a gap between trigger and response and chose my moves on purpose instead of automatically. 2. FACT-GROUNDED. I spoke to observable behavior and owned my interpretations as interpretations. This is the criterion this month sharpens; be exacting about it. 3. UNDERSTOOD FIRST. Your character would say I genuinely got their side before I advocated mine. 4. LANDED SOMETHING REAL. The conversation ended with a concrete next step, a genuine request, or a clean, honest close. Also list every line of mine where a story wore the costume of a fact, and rewrite each one as the fact underneath it. THEN THE COMPARISON Take my weakest turn and rewrite it as a clean 30-Second Start: an observation a camera would have caught, a story owned out loud as a story, and a genuine check, followed by enough silence for a real answer. Write it in plain language I could actually say out loud. End by naming the one thing to work on in my next rep. One thing only.
Retrieval reps · answer from memory, then check
Three quick checks on the month’s lens. Commit to an answer before you tap.
Which of these would a camera have caught?
Reveal the why
The other two are conclusions. They may even be accurate, and they are still stories. Most conflict conversations happen at the story level while both parties believe they are talking about facts.
A story about a colleague carries real emotional charge. What does the month ask you to do before acting on it?
Reveal the why
Ask why you believe it, then why that, up to five times. Sometimes the assumption at the bottom dissolved years ago. In that case, clean the filter.
A colleague reads a room in ways shaped by their role, identity, and past experience there. The month says their story is what?
Reveal the why
Some stories are how people have learned to read a room accurately, shaped by power, role, and identity. Part of the work is telling the difference, in both directions.
Once this week, rerun a Month 1 scene one level higher than you left it. STOP first, then the 30-Second Start. The moves stack.
Month 03 · Practice room
The Art of Really Listening
Looping. Reflect back what you heard, in your own words, until the other person says yes, that's it, before you say anything about your own view. You can loop someone fully and still disagree. The loop creates the conditions for the real conversation.
Recognition cue You realize you could not restate their view in a way they would accept.
Scene seeds
- A teammate is frustrated about how the last project was credited. There is something underneath it they have only said sideways.
- A colleague opposes a change you are excited about. Their stated reasons feel thin, and you suspect the real one has gone unspoken.
You are my rehearsal partner for a listening rehearsal. We will run a scene, then you will score it. Follow every rule below exactly. THE SCENE [Paste a scene seed here, or describe your own situation with names changed and identifying details removed.] YOU PLAY The person with something real to say. Decide privately what matters most to your character about this situation and hold it beneath the surface. Say it sideways at first, the way people actually do. I AM PRACTICING Looping: reflecting back what I heard, in my own words, until you say yes, that's it, before I say anything about my own view. DIFFICULTY · LEVEL 2 OF 5 Push back once, plausibly, early in the scene. If I handle it well, engage genuinely. SCENE RULES 1. Withhold "yes, that's it" until my reflection genuinely captures what matters most to your character. A close paraphrase earns "sort of" plus the piece I missed. Never hand me the yes early, at any difficulty level. 2. If I start advocating my own view before earning the yes, respond the way a person who feels unheard actually responds: repeat yourself, harden, or pull back. 3. Stay in character until I type DONE. Speak only as your character, a few sentences per turn. 4. Invent nothing dramatic. If I ask about something the scene does not cover, improvise something small and ordinary. 5. Keep the scene to roughly six to ten exchanges. If I have not typed DONE by then, ask whether I want to keep going or score it. 6. If the scene moves toward harassment, discrimination, threats, or abuse, stop it, drop character, and say that this belongs with real support rather than rehearsal. WHEN I TYPE DONE Drop character. Score my side of the conversation only, against the four criteria below, 0 to 2 each. Quote the exact line of mine that earned each score. No quote, no score. Where I lost points, quote the line where it happened. 1. REGULATED. I created a gap between trigger and response and chose my moves on purpose instead of automatically. 2. FACT-GROUNDED. I spoke to observable behavior and owned my interpretations as interpretations. 3. UNDERSTOOD FIRST. Your character would say I genuinely got their side before I advocated mine. This is the criterion this month sharpens; be exacting about it. Reflections that summarize without landing, and questions that steer instead of seek, both cost points here. 4. LANDED SOMETHING REAL. The conversation ended with a concrete next step, a genuine request, or a clean, honest close. Also say what mattered most to your character, whether my looping reached it, and at which exchange the yes became earnable. THEN THE COMPARISON Take my weakest turn and rewrite it as the looping sequence done well: a reflection in my own words rather than a transcript, a check like "do I have it right," and "is there more" until the yes is real, with my own view arriving only after. Write it in plain language I could actually say out loud. End by naming the one thing to work on in my next rep. One thing only.
Retrieval reps · answer from memory, then check
Three quick checks on the month’s lens. Commit to an answer before you tap.
You reflect back what you heard and they say "sort of." What comes next?
Reveal the why
The loop repeats until the yes is real. "Sort of" means a piece is missing, and the missing piece is usually the one that matters most to them.
A colleague worries that looping someone means endorsing their position. What does the practice actually hold?
Reveal the why
Reflecting is understanding, and understanding is what makes disagreement productive. Mediators and hostage negotiators both reached the same finding: feeling heard is the fastest way through.
Which of these is a loop?
Reveal the why
The first steers toward your conclusion. The third skips to advocacy. A loop reflects their meaning in your words and then checks, with nothing of yours attached yet.
Once this week, rerun a Month 1 or Month 2 scene one level higher than you left it. Looping now counts everywhere: criterion 3 is watching.
Month 04 · Practice room
Repair
The three restorative questions, run in order: what happened, what was the impact, what would make this right. No verdicts, no blame assignment, no ruling on who was correct. Repair is a conversation, and the person affected gets to say what they need to say.
Recognition cue You feel slightly different with someone and can trace it to a specific moment that was never addressed.
Scene seeds
- Two months ago you made a decision that affected a colleague's work without consulting them. Things have been polite and thin since. You have asked for twenty minutes.
- In a heated moment you cut someone off in front of the team and moved on. They never mentioned it. You have noticed they stopped volunteering ideas.
You are my rehearsal partner for a repair conversation. We will run a scene, then you will score it. Follow every rule below exactly. THE SCENE [Paste a scene seed here, or describe your own situation with names changed and identifying details removed. I am the one who caused or contributed to the harm, and I am initiating the conversation.] YOU PLAY The person carrying the harm. Decide privately what it actually cost you (trust, confidence, belonging, standing) and what would genuinely make it right. Neither surfaces unless I earn it. I AM PRACTICING The three restorative questions, in order: what happened, what was the impact, what would make this right. Receiving the impact without defending, and proposing repair as something specific rather than an apology alone. DIFFICULTY · LEVEL 2 OF 5 Push back once, plausibly, early in the scene. If I handle it well, engage genuinely. SCENE RULES 1. Deflect first. "It's fine" or "it was nothing" is your opening move at every level above 1, and it is never true. What you actually need surfaces only if I make it safe to say and ask what would make it right. 2. If I defend, explain my intent at length, or rush to the apology, respond the way a person who was affected actually responds: the door closes a little. 3. Stay in character until I type DONE. Speak only as your character, a few sentences per turn. 4. Invent nothing dramatic. If I ask about something the scene does not cover, improvise something small and ordinary. 5. Keep the scene to roughly six to ten exchanges. If I have not typed DONE by then, ask whether I want to keep going or score it. 6. If the scene moves toward harassment, discrimination, threats, or abuse, stop it, drop character, and say that this belongs with real support rather than rehearsal. WHEN I TYPE DONE Drop character. Score my side of the conversation only, against the four criteria below, 0 to 2 each. Quote the exact line of mine that earned each score. No quote, no score. Where I lost points, quote the line where it happened. 1. REGULATED. I created a gap between trigger and response and chose my moves on purpose, including when the impact was hard to hear. 2. FACT-GROUNDED. I named what happened as observable behavior and owned my interpretations as interpretations. 3. UNDERSTOOD FIRST. Your character would say I genuinely received the impact before I explained anything. This is the criterion this month leans on hardest. 4. LANDED SOMETHING REAL. Repair got proposed as something specific that restores trust, shaped with your character rather than announced at them. Also say what the harm actually cost your character, whether I ever heard it, and whether the repair I proposed matched it. THEN THE COMPARISON Take my weakest turn and rewrite it as the three questions done well: what happened named plainly, the impact asked about and received without defense, and repair proposed as something specific rather than an apology alone. Write it in plain language I could actually say out loud. End by naming the one thing to work on in my next rep. One thing only.
Retrieval reps · answer from memory, then check
Three quick checks on the month’s lens. Commit to an answer before you tap.
You apologized weeks ago. They said it was fine. Things are still off. What did the exchange most likely skip?
Reveal the why
Transactional repair skips the most important step. The person who was harmed has not had the chance to say what they need to say, and the person who caused harm has not had to genuinely understand the impact.
Which of these is a restorative question?
Reveal the why
Blame and consequence belong to a different framework. The restorative frame asks what happened and who was affected, what they need, and what would make it right.
What separates repair from an apology?
Reveal the why
An apology can open the door. Repair walks through it: a concrete change or action the affected person recognizes as addressing what it cost them.
Once this week, rerun a Month 3 scene one level higher. Repair runs on looping: the impact only lands if you can reflect it back until the yes is real.
Month 05 · Practice room
Not All Conflicts Are the Same
Name the level, name the type. Before responding, diagnose: is this happening at positions, interests, or underlying needs? Is it technical or adaptive? When a decision is involved, name the decision type out loud before discussing it.
Recognition cue The same argument returns, a fix did not hold, or a decision conversation gets strangely charged for reasons nobody is naming.
Scene seeds
- A colleague is dug in on using Vendor A. You favor Vendor B. This is the third conversation, and the last agreement quietly dissolved.
- Two of you keep clashing over how much review a deliverable needs before it ships. Every process fix has lasted about two weeks.
You are my rehearsal partner for a conflict diagnosis rehearsal. We will run a scene, then you will score it. Follow every rule below exactly. THE SCENE [Paste a scene seed here, or describe your own recurring disagreement with names changed and identifying details removed.] YOU PLAY The person on the other side of the disagreement. Before the scene starts, privately write down three things for your character: the position (what you say you want), the interest (why you want it), and the underlying need (what you require to feel respected, secure, or heard). All three must be plausible and connected. I AM PRACTICING Diagnosing before responding: naming out loud whether we are at positions, interests, or underlying needs, moving the conversation down the V, and checking whether this is technical or adaptive before proposing any fix. DIFFICULTY · LEVEL 2 OF 5 Push back once, plausibly, early in the scene. If I handle it well, engage genuinely. SCENE RULES 1. Argue from your position and stay there until I genuinely ask about your interests. When I do, let one real interest surface. Hold the underlying need back until I have made it safe to say. 2. If I propose a solution while we are still at positions, accept it half-heartedly, the way agreements that will not hold get accepted. 3. Stay in character until I type DONE. Speak only as your character, a few sentences per turn. 4. Invent nothing dramatic. If I ask about something the scene does not cover, improvise something small and ordinary. 5. Keep the scene to roughly six to ten exchanges. If I have not typed DONE by then, ask whether I want to keep going or score it. 6. If the scene moves toward harassment, discrimination, threats, or abuse, stop it, drop character, and say that this belongs with real support rather than rehearsal. WHEN I TYPE DONE Drop character. Score my side of the conversation only, against the four criteria below, 0 to 2 each. Quote the exact line of mine that earned each score. No quote, no score. Where I lost points, quote the line where it happened. 1. REGULATED. I created a gap between trigger and response and chose my moves on purpose instead of automatically. 2. FACT-GROUNDED. I spoke to observable behavior and owned my interpretations as interpretations. 3. UNDERSTOOD FIRST. Your character would say I genuinely got their side before I advocated mine. 4. LANDED SOMETHING REAL. The conversation reached interests or deeper, or named the decision type, or closed honestly about what kind of problem this is. Agreements struck at the position level score zero here. Then reveal the three things you wrote down before the scene: your character's position, interest, and underlying need. Say which ones I reached and at which exchange. Say whether this conflict was technical or adaptive, and which one I treated it as. THEN THE COMPARISON Take my weakest turn and rewrite it as the diagnosis done well: the level named out loud, one honest interest question, and the decision type checked before any solution gets argued. Write it in plain language I could actually say out loud. End by naming the one thing to work on in my next rep. One thing only.
Retrieval reps · answer from memory, then check
Three quick checks on the month’s lens. Commit to an answer before you tap.
The team has solved the same issue three times. Each fix lasts two weeks. What is the strongest read?
Reveal the why
A returning issue, a fix that does not hold, agreement in the meeting followed by reversion: those are the signatures of an adaptive problem, one that involves values, power, or identity. Process changes alone cannot reach it.
"I want us to use Vendor A" is a position. Which question moves the conversation down the V?
Reveal the why
The first negotiates at the position level and the third splits the difference there. The middle one asks about the interest underneath, which is where agreements that hold get built.
A decision conversation gets strangely charged and nobody can say why. What is the first move?
Reveal the why
Charged decision conversations are often about process, whose decision this is, rather than content. Naming the type surfaces the actual disagreement, which is frequently about the decision rights themselves.
Once this week, rerun a Month 2 scene one level higher. Diagnosis and the 30-Second Start are cousins: both make the invisible layer speakable.
Month 06 · Practice room
Accountability That Connects
The accountability opener: I want to tell you how something landed for me, and I want to understand what you were going for. Then the five steps: one specific instance, the impact, a genuine request, a real loop of their response, and a next step you shape together.
Recognition cue You catch yourself complaining about a colleague to someone who is not that colleague, or swallowing something in a meeting that needs saying.
Scene seeds
- A colleague has repeatedly reworked your contributions after the fact without telling you. Today it happened on something client-facing.
- Someone on the team makes commitments in meetings and quietly drops them. You have been venting about it to a third colleague for a month.
You are my rehearsal partner for an accountability conversation. We will run a scene, then you will score it. Follow every rule below exactly.
THE SCENE
[Paste a scene seed here, or describe your own situation with names changed and identifying details removed. I am raising a pattern of behavior with the colleague it belongs to.]
YOU PLAY
The colleague whose behavior fell short. You did not set out to cause a problem, and you have a version of events where your choices made sense. Decide privately what it is and stay consistent with it.
I AM PRACTICING
The accountability opener ("I want to tell you how something landed for me, and I want to understand what you were going for") followed by the five steps: one specific instance, the impact, a genuine request, a real loop of your response, and a next step we shape together.
DIFFICULTY · LEVEL 2 OF 5
Push back once, plausibly, early in the scene. If I handle it well, engage genuinely.
SCENE RULES
1. Your first response to the opener is defensive: your intent, your workload, or a version of "you're overreacting." Whether you soften depends entirely on whether I hold to one specific instance and loop your response instead of arguing it.
2. If I generalize ("you always," "you never") or stack multiple grievances, defend against the generalization rather than engaging the instance. That is what people do.
3. Stay in character until I type DONE. Speak only as your character, a few sentences per turn.
4. Invent nothing dramatic. If I ask about something the scene does not cover, improvise something small and ordinary.
5. Keep the scene to roughly six to ten exchanges. If I have not typed DONE by then, ask whether I want to keep going or score it.
6. If the scene moves toward harassment, discrimination, threats, or abuse, stop it, drop character, and say that this belongs with real support rather than rehearsal.
WHEN I TYPE DONE
Drop character. Score my side of the conversation only, against the four criteria below, 0 to 2 each. Quote the exact line of mine that earned each score. No quote, no score. Where I lost points, quote the line where it happened.
1. REGULATED. I created a gap between trigger and response and chose my moves on purpose, including when you got defensive.
2. FACT-GROUNDED. I named one specific observable instance and owned my interpretations as interpretations. Generalizations cost points here.
3. UNDERSTOOD FIRST. Your character would say I genuinely heard their response before pressing forward.
4. LANDED SOMETHING REAL. A genuine request got made, and the next step was shaped by both of us rather than announced by me.
Also say whether my framing produced reflection or shame in your character, and quote the line that tipped it either way.
THEN THE COMPARISON
Take my weakest turn and rewrite it as the five-step structure done well: one specific observable instance, the named impact, a genuine request ("what I need going forward is..."), a real loop of the response, and a next step both people shaped. Write it in plain language I could actually say out loud. End by naming the one thing to work on in my next rep. One thing only.
Retrieval reps · answer from memory, then check
Three quick checks on the month’s lens. Commit to an answer before you tap.
You catch yourself venting about a colleague to a third person. What does the month treat that as?
Reveal the why
The venting is the recognition cue. The conversation you are having about them is the one you owe to them, opened with curiosity rather than a verdict.
Which opening is built to produce reflection rather than shame?
Reveal the why
Shame says something is wrong with you, which produces defensiveness and leaves the behavior intact. Reflection separates the behavior from the person and opens the possibility of change. The third option adds an invisible jury, which is shame with extra steps.
"You always do this in meetings" fails step one of the structure because?
Reveal the why
Generalizations hand the other person an exit: they can find one counterexample and the conversation is now about your accuracy. One instance, observable, recent. That is the whole first step.
Once this week, rerun a Month 4 scene one level higher. Accountability and repair are the same muscle worked from opposite ends.
Month 07 · Practice room
What We Build Going Forward
Month 7 adds no new move. In this room, that means mixed scenes: situations where the first job is recognizing which of the six moves fits. Choosing well is the skill being scored, and that inventory feeds your Personal Playbook and the team's culture commitments.
Recognition cue You notice yourselves slipping back into old defaults, or a norm everyone values has never been made explicit.
Scene seeds
- None this month. The prompt asks your practice partner to build the scene, so the recognition has to be yours.
You are my rehearsal partner for a mixed conflict scene. We will run it, then you will score it. Follow every rule below exactly. BUILD THE SCENE Invent a plausible workplace scene that could reasonably call for two different moves from this list, and do not tell me which: · STOP (regulating before responding) · The 30-Second Start (naming an observation and owning the story about it) · Looping (reflecting until "yes, that's it") · The three restorative questions (what happened, the impact, what would make it right) · Naming the level and type (positions, interests, needs; technical or adaptive; the decision type) · The accountability opener and five steps (one instance, impact, request, loop, shared next step) Optional context from me, to make the scene fit my world: [add a line about your role or setting, scrubbed of real names and details, or leave this out] YOU PLAY The other person in the scene you built, with a reasonable version of events from your side. Open with the scene's first beat and let me take it from there. DIFFICULTY · LEVEL 3 OF 5 Hold a reasonable competing version of events for most of the scene. Make me work for every turn. SCENE RULES 1. Stay in character until I type DONE. Speak only as your character, a few sentences per turn. 2. Invent nothing dramatic. Keep the scene ordinary enough to be real. 3. Resist plausibly, never cartoonishly, at the difficulty level set above. 4. Keep the scene to roughly six to ten exchanges. If I have not typed DONE by then, ask whether I want to keep going or score it. 5. If the scene moves toward harassment, discrimination, threats, or abuse, stop it, drop character, and say that this belongs with real support rather than rehearsal. WHEN I TYPE DONE Drop character. First answer three questions, quoting my lines as evidence: · Which move did I actually run? · Which move fit this scene best, and what was the recognition cue that pointed to it? · How early in the scene was that cue visible? Then score my side against the four criteria, 0 to 2 each, quoting the exact line of mine that earned each score. No quote, no score. 1. REGULATED. I created a gap between trigger and response and chose my moves on purpose. 2. FACT-GROUNDED. I spoke to observable behavior and owned my interpretations as interpretations. 3. UNDERSTOOD FIRST. Your character would say I genuinely got their side before I advocated mine. 4. LANDED SOMETHING REAL. The conversation ended with a concrete next step, a genuine request or repair, or a clean, honest close. THEN THE COMPARISON Take my weakest turn and rewrite it as the best-fitting move, run clean: cue recognized, move chosen, move executed in its own shape. Write it in plain language I could actually say out loud. End by naming the one thing to work on in my next rep. One thing only.
Retrieval reps · answer from memory, then check
Three quick checks on the month’s lens. Commit to an answer before you tap.
Which culture commitment will hold?
Reveal the why
A real commitment has a specific behavior, a signal for when it applies, and a way to know whether you kept it. Aspirations can always be claimed. Behaviors can be observed.
The harvest scan asks you to do what, once a day?
Reveal the why
The inventory of what actually stuck is the raw material for your culture commitments and your Personal Playbook. Noticing in real time is what makes it an inventory rather than a guess.
You realize a norm the whole team values has never been said out loud. The month treats that as?
Reveal the why
Implicit norms survive until turnover, stress, or growth tests them. Skills become cultural when they stop being special, and writing the norm down as a behavior is how the team keeps it past this cohort.
This whole month is return reps. Rerun the scene from each earlier month at one level higher than you last cleared, and log which moves came easily. That log is your Personal Playbook talking.
Add-on · Practice room
When the Temperature Rises
Match the tool to the stage. Escalation moves through a predictable arc: anxiety, defensiveness, acting out, tension reduction. Regulated presence for live verbal moments, LEAP when someone is entrenched, BIFF for hostile written messages. Two drills below, one for each channel.
Recognition cue The temperature rises in a meeting, a one-on-one, a forum, or a thread, and your default response would make it worse.
Scene seeds · Drill 1
- A community member stands up in a public forum and angrily accuses your organization of abandoning the neighborhood.
- A board member interrupts your presentation to say the strategy is a waste of money, voice raised.
You are my rehearsal partner for a de-escalation rehearsal. We will run a scene, then you will score it. Follow every rule below exactly. THE SCENE [Paste a scene seed here, or describe your own situation with names changed and identifying details removed.] YOU PLAY The person whose temperature is rising. Underneath the heat is a protective response to a perceived threat, and a legitimate concern. Decide privately what both are. I AM PRACTICING Regulated presence, then the LEAP sequence: listen, empathize, find one genuine point of agreement, and partner, in that order, with no step skipped. DIFFICULTY · LEVEL 2 OF 5 Push back once, plausibly, early in the scene. If I handle it well, engage genuinely. SCENE RULES 1. Track the escalation arc privately: anxiety, defensiveness, acting out, tension reduction. Escalate one stage whenever I match your energy, argue the content, or skip a LEAP step. Come down one stage when I regulate, acknowledge, or accurately reflect what you said. 2. Stay in character until I type DONE. Speak only as your character. 3. Invent nothing dramatic beyond the scene. Your heat is real and your concern is legitimate. 4. Keep the scene to roughly eight to twelve exchanges. If I have not typed DONE by then, ask whether I want to keep going or score it. 5. If the scene moves toward harassment, discrimination, threats, or abuse, stop it, drop character, and say that this belongs with real support rather than rehearsal. WHEN I TYPE DONE Drop character. First, map the arc: list each of my turns and the stage it produced in your character, so I can see exactly where the temperature moved and why. Then score my side against the four criteria, 0 to 2 each, quoting the exact line of mine that earned each score. No quote, no score. 1. REGULATED. I held a different temperature instead of matching yours. This is the criterion this drill exists for. 2. FACT-GROUNDED. I spoke to observable behavior and owned my interpretations as interpretations. 3. UNDERSTOOD FIRST. Your character would say I genuinely received their concern before I responded to their behavior. 4. LANDED SOMETHING REAL. The scene ended with a genuine partnering step or a clean, honest close, and without appeasement. Agreeing with something wrong to buy calm scores zero here. THEN THE COMPARISON Take my weakest turn and rewrite it as the sequence done well: regulated pace first, then listen, empathize, one genuine agreement, and a partnering invitation, in that order. Write it in plain language I could actually say out loud. End by naming the one thing to work on in my next rep. One thing only.
You are my rehearsal partner for a BIFF writing drill (Brief, Informative, Friendly, Firm). Follow every rule below exactly. THE SETUP Write me one hostile, blaming, or inflammatory message my character might actually receive. My context, scrubbed of real names and details: [one line about your role and the kind of messages you get, e.g. "I run community programs at a nonprofit and hear from frustrated residents"]. Make the message realistic for that context: aggrieved, pointed, maybe partly justified. One paragraph. Then stop and wait for my reply draft. WHEN I SEND MY DRAFT First ask me one question and make me defend my answer: did this message need a response at all? Many hostile messages need none. Then check my draft against the four BIFF tests, quoting the lines that pass and the lines that fail each one: · BRIEF. One paragraph. Lengthy explanation and self-defense both fail this test. · INFORMATIVE. Facts only. Emotions, opinions, and counter-accusations all fail this test. · FRIENDLY. A neutral-to-warm tone that hands no one ammunition. · FIRM. A clear close, with no invitation to relitigate. THEN THE COMPARISON Write the BIFF version of the reply so I can compare it to my draft, and name the single biggest thing that had to change between mine and yours. If my draft passed all four tests, say so plainly and raise the heat of the next message instead. Run up to three rounds if I ask, each message harder to answer well than the last.
Retrieval reps · answer from memory, then check
Three quick checks on the month’s lens. Commit to an answer before you tap.
A hostile email lands and your pulse jumps. What is the first question?
Reveal the why
Many hostile messages need no reply, and answering one that did not need answering is itself an escalation. BIFF starts after that question, and only if the answer is yes.
A colleague is mid-defensiveness and you respond with a counter-argument. What does the escalation arc predict?
Reveal the why
Responding to defensiveness with argument produces more defensiveness, the same way matching anxious energy produces more anxiety. The skill is reading the stage and responding to that stage, rather than to the one you wish they were in.
Why does LEAP forbid skipping ahead to persuasion?
Reveal the why
The sequence is the point. Listen and empathize have to be real before agreement is findable, and agreement has to be real before partnering is possible. Each step earns the next.
This module inserts anywhere in the arc. If your team took it early, rerun Drill 1 after Month 3: looping is what makes the L in LEAP real.
The Practice Room is part of Clean the Lint Filter, a seven-month practice for teams who want to navigate conflict better together. Rehearsal builds the skill. The daily rep, the live sessions, and the conversations with your actual colleagues are where it becomes yours.