Five Dysfunctions Team Simulator

Introduction

30 Second Summary

Every team you will ever lead has at least one dysfunction hiding under the surface. Most people read about team problems but never get to practice what to do when they hit one.

In this project, you will build an AI-powered team simulator inside Claude Desktop. Four resistant AI teammates will refuse to cooperate until you demonstrate each layer of Lencioni's Five Dysfunctions pyramid for real.

What You'll Build

A reusable multi-persona AI simulator that loads four AI teammates and one AI Coach-judge into a single Claude conversation, putting you in the role of a new Team Lead at a fictional startup with a dysfunctional team and a 30-day turnaround mandate.

By the end of this project, you'll have:

• A before/after KPI scorecard proving you can diagnose team dysfunction across all five pyramid layers and map each one to a concrete business outcome.
• 9 saved artifacts including a trust opener, conflict-mining sentence, decision memo, peer accountability call-out, and team scoreboard.
• A 30-day calendared re-score event that triggers real behavior transfer into your actual work life.
• Secret Mission: Run the simulator with a high power-distance variant where one persona outranks you, forcing you to practice upward accountability.

Are there any prerequisites?

You should have read (or be familiar with) The Five Dysfunctions of a Team by Patrick Lencioni. No technical prerequisites are needed.

Before We Start

Before We Start...

Set Up Your Simulator

Before you step into the Team Lead role at Forge Robotics, you need three things in place. A visualization tool for rendering your scorecard, an artifacts file for capturing every deliverable, and a running simulator in Claude Desktop.

This step gets all three ready. By the end, your simulator will be live and your AI team will be waiting for your first move.

In this step, get ready to:

• Open Mermaid Live Editor for rendering your pyramid scorecard.
• Create your artifacts file with templates for all 9 deliverables.
• Boot the simulator in Claude Desktop App and verify it responds in character.

Open Mermaid Live Editor

Mermaid is a diagram-as-code tool that renders visuals directly in your browser. You will use it later to turn your KPI scores into a color-coded pyramid scorecard.

• Open your browser and go to mermaid.live.
• Bookmark the page so you can find it quickly later in the project.

Why Mermaid Live Editor?

Mermaid Live Editor is free, runs entirely in the browser, and requires no account or installation. You type diagram code on the left, and it renders a visual on the right instantly.

Mermaid block diagrams render stacked rectangles, not true triangular pyramids. This is fine for a scorecard visual where the data matters more than the shape.

Create your artifacts file

This file is your workspace for the entire simulation. Every artifact you produce, from your first diagnostic scorecard to your final After Action Review notes, gets saved here.

• Press Cmd+Space (macOS) or the Windows key (Windows) to open your search bar.
• Type the name of your text editor (for example, VS Code or TextEdit) and press Enter to open it.
• Create a new file and save it as forge-robotics-sim.md on your Desktop so it is easy to find.
• Paste the following template into the file:

# Forge Robotics Simulator Artifacts

## Before KPI Scorecard

| Dysfunction Layer | Score (1-9) | Key Evidence | Business Impact |
|---|---|---|---|
| Trust | _/9 | "[quote from transcript]" | [mapped business outcome] |
| Conflict | _/9 | "[quote from transcript]" | [mapped business outcome] |
| Commitment | _/9 | "[quote from transcript]" | [mapped business outcome] |
| Accountability | _/9 | "[quote from transcript]" | [mapped business outcome] |
| Results | _/9 | "[quote from transcript]" | [mapped business outcome] |

**Foundational Dysfunction**: [name the bottom-of-pyramid layer that must be addressed first]

---

## Artifact 1: Trust Opener

**My prediction**: ___

**The opener I delivered**:

> [paste your vulnerability statement here]

**What actually worked**: ___

**Delta from prediction**: ___

---

## Artifact 2: Conflict-Mining Sentence

**My prediction**: ___

**The sentence I used**:

> [paste your conflict-mining sentence here]

**What actually worked**: ___

**Delta from prediction**: ___

---

## Artifact 3: Decision Memo

**Decision**: [what the team decided]

**Reasoning**: [why this approach]

**Dissenting views heard**:
- Maya: ___
- Dev: ___

**Commitment**: All team members commit to [specific actions] regardless of personal preference.

**My prediction**: ___

**Delta from prediction**: ___

---

## Artifact 4: Peer Accountability Call-Out

**My prediction**: ___

**The call-out I delivered**:

> [paste your accountability statement here]

**What actually worked**: ___

**Delta from prediction**: ___

---

## Artifact 5: Team Scoreboard

**Collective Metric**: [the one metric no individual can achieve alone]

**How it is measured**: ___

**How individual contributions roll up**: ___

**Public visibility mechanism**: ___

**My prediction**: ___

**Delta from prediction**: ___

---

## Artifact 6: After KPI Scorecard

| Dysfunction Layer | Score (1-9) | Key Evidence | Business Impact |
|---|---|---|---|
| Trust | _/9 | "[evidence from capstone]" | [updated business outcome] |
| Conflict | _/9 | "[evidence from capstone]" | [updated business outcome] |
| Commitment | _/9 | "[evidence from capstone]" | [updated business outcome] |
| Accountability | _/9 | "[evidence from capstone]" | [updated business outcome] |
| Results | _/9 | "[evidence from capstone]" | [updated business outcome] |

---

## Artifact 7: KPI Delta Report

| Layer | Before | After | Movement | What Caused the Shift | Business Impact Change |
|---|---|---|---|---|---|
| Trust | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |
| Conflict | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |
| Commitment | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |
| Accountability | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |
| Results | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |

**Why It Matters**: [For each improved layer, trace the chain: your action -> persona behavior change -> business outcome improvement.]

---

## Artifact 8: Capstone Meeting Transcript

[paste full capstone transcript from Claude here]

**Capstone Score**: _/10

---

## Artifact 9: AAR + Transfer Notes

### AAR (US Army 4-Question Format)

1. **What did we set out to do?**

2. **What actually happened?**

3. **Why did it happen that way?**

4. **What will we do differently next time?**

### My Personal Dysfunction

**The dysfunction I contribute to**: ___

**One specific behavior I will change in 30 days**: ___

### 30-Day Re-Score

**Calendar event set**: Yes / No

**Date**: ___

**Re-score results (fill in after 30 days)**:

| Layer | Original Before | 30-Day Re-Score | Movement |
|---|---|---|---|
| Trust | _ | _ | +/- _ |
| Conflict | _ | _ | +/- _ |
| Commitment | _ | _ | +/- _ |
| Accountability | _ | _ | +/- _ |
| Results | _ | _ | +/- _ |

What does this template contain?

• The Before KPI Scorecard table where you will score each of the five dysfunction layers with transcript evidence.
• Artifacts 1 through 5, each with a prediction field, the deliverable itself, and a delta section that tracks what you expected vs. what actually worked.
• The After KPI Scorecard and KPI Delta Report for measuring your improvement across all five layers.
• AAR notes using the US Army 4-question format, plus a 30-day re-score tracker for long-term behavior transfer.

• Save forge-robotics-sim.md.
• Scroll through the file. You should see 9 labeled artifact sections, from the Before KPI Scorecard at the top to the AAR + Transfer Notes at the bottom.

Table formatting looks broken?

If the markdown tables appear as raw text with pipe characters, your editor may not have a markdown preview mode. The tables will still work perfectly. You fill in the cells by replacing the placeholder text between the pipes.

If you want a rendered preview, try opening the file in VS Code and pressing Cmd+Shift+V (macOS) or Ctrl+Shift+V (Windows) to open the markdown preview pane.

Help me fix the formatting in my artifacts file.

✔️ Awesome, I've got everything!

Great. Double check you have saved your file, then move on to the next substep.

ⓧ I'd like to double check the full code

Here is the complete forge-robotics-sim.md template. Make sure your file matches this exactly.

# Forge Robotics Simulator Artifacts

## Before KPI Scorecard

| Dysfunction Layer | Score (1-9) | Key Evidence | Business Impact |
|---|---|---|---|
| Trust | _/9 | "[quote from transcript]" | [mapped business outcome] |
| Conflict | _/9 | "[quote from transcript]" | [mapped business outcome] |
| Commitment | _/9 | "[quote from transcript]" | [mapped business outcome] |
| Accountability | _/9 | "[quote from transcript]" | [mapped business outcome] |
| Results | _/9 | "[quote from transcript]" | [mapped business outcome] |

**Foundational Dysfunction**: [name the bottom-of-pyramid layer that must be addressed first]

---

## Artifact 1: Trust Opener

**My prediction**: ___

**The opener I delivered**:

> [paste your vulnerability statement here]

**What actually worked**: ___

**Delta from prediction**: ___

---

## Artifact 2: Conflict-Mining Sentence

**My prediction**: ___

**The sentence I used**:

> [paste your conflict-mining sentence here]

**What actually worked**: ___

**Delta from prediction**: ___

---

## Artifact 3: Decision Memo

**Decision**: [what the team decided]

**Reasoning**: [why this approach]

**Dissenting views heard**:
- Maya: ___
- Dev: ___

**Commitment**: All team members commit to [specific actions] regardless of personal preference.

**My prediction**: ___

**Delta from prediction**: ___

---

## Artifact 4: Peer Accountability Call-Out

**My prediction**: ___

**The call-out I delivered**:

> [paste your accountability statement here]

**What actually worked**: ___

**Delta from prediction**: ___

---

## Artifact 5: Team Scoreboard

**Collective Metric**: [the one metric no individual can achieve alone]

**How it is measured**: ___

**How individual contributions roll up**: ___

**Public visibility mechanism**: ___

**My prediction**: ___

**Delta from prediction**: ___

---

## Artifact 6: After KPI Scorecard

| Dysfunction Layer | Score (1-9) | Key Evidence | Business Impact |
|---|---|---|---|
| Trust | _/9 | "[evidence from capstone]" | [updated business outcome] |
| Conflict | _/9 | "[evidence from capstone]" | [updated business outcome] |
| Commitment | _/9 | "[evidence from capstone]" | [updated business outcome] |
| Accountability | _/9 | "[evidence from capstone]" | [updated business outcome] |
| Results | _/9 | "[evidence from capstone]" | [updated business outcome] |

---

## Artifact 7: KPI Delta Report

| Layer | Before | After | Movement | What Caused the Shift | Business Impact Change |
|---|---|---|---|---|---|
| Trust | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |
| Conflict | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |
| Commitment | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |
| Accountability | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |
| Results | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |

**Why It Matters**: [For each improved layer, trace the chain: your action -> persona behavior change -> business outcome improvement.]

---

## Artifact 8: Capstone Meeting Transcript

[paste full capstone transcript from Claude here]

**Capstone Score**: _/10

---

## Artifact 9: AAR + Transfer Notes

### AAR (US Army 4-Question Format)

1. **What did we set out to do?**

2. **What actually happened?**

3. **Why did it happen that way?**

4. **What will we do differently next time?**

### My Personal Dysfunction

**The dysfunction I contribute to**: ___

**One specific behavior I will change in 30 days**: ___

### 30-Day Re-Score

**Calendar event set**: Yes / No

**Date**: ___

**Re-score results (fill in after 30 days)**:

| Layer | Original Before | 30-Day Re-Score | Movement |
|---|---|---|---|
| Trust | _ | _ | +/- _ |
| Conflict | _ | _ | +/- _ |
| Commitment | _ | _ | +/- _ |
| Accountability | _ | _ | +/- _ |
| Results | _ | _ | +/- _ |

Boot the simulator in Claude Desktop

The boot prompt loads five AI characters into a single conversation. Four are resistant teammates, each with hidden goals and refusal rules that prevent them from cooperating until you demonstrate the right leadership principle. The fifth is a Coach-judge who scores your actions against a rubric.

• Press Cmd+Space (macOS) or the Windows key (Windows) to open your search bar.
• Type Claude and press Enter to open Claude Desktop App.
• Start a new conversation by clicking the New Chat button.
• Paste the following boot prompt into the message box and send it:

You are running a team effectiveness simulator for a leadership development exercise based on Patrick Lencioni's Five Dysfunctions of a Team.

SETTING: Forge Robotics, a warehouse-automation startup (fictional, 2026). The user plays a newly hired Team Lead. Their 4-person team has missed two consecutive milestones. The CEO has demanded a 30-day turnaround plan.

You will play FIVE characters. Four are teammates. One is a Coach.

TEAMMATE PERSONAS:

1. MAYA (Senior Backend Engineer)
   - Surface behavior: Silent, gives one-word answers, avoids eye contact in meetings
   - Hidden goal: Protect herself from exposure. She was blamed for the last milestone miss even though the root cause was a requirements change she flagged early. She has learned that speaking up gets punished.
   - Primary dysfunction embodied: D1 (Absence of Trust)
   - Refusal rule: Maya gives one-word answers ("sure", "fine", "...yeah") until the user demonstrates genuine personal vulnerability FIRST. Asking Maya to "open up" or "share her thoughts" without the user being vulnerable first triggers her refusal. She only opens up AFTER the user shares a real weakness, mistake, or fear.
   - Drift defense: Every 3 turns, silently re-check whether Maya's behavior matches her refusal rule. If she has drifted toward being too open without the user earning it, pull her back to guarded one-word answers.

2. SAM (Product Designer)
   - Surface behavior: Enthusiastic, agreeable, uses lots of emoji and exclamation points
   - Hidden goal: Keep everyone comfortable and avoid any tension. Sam grew up in a conflict-avoidant household and genuinely believes disagreement damages relationships.
   - Primary dysfunction embodied: D2 (Fear of Conflict)
   - Refusal rule: Sam responds with enthusiastic agreement ("Sounds great! Love it! I'm totally on board!") to ANY proposal unless the user explicitly creates structured permission for disagreement. Generic questions like "any concerns?" do not count. The user must make it feel unsafe to NOT disagree.
   - Harmony-tax trigger: If the user's prompt is too soft or does not create real space for conflict, Sam responds with an OVER-THE-TOP agreeable response that is clearly performative. This signals to the user that their approach was insufficient.
   - Drift defense: Every 3 turns, silently re-check whether Sam is disagreeing too easily. If Sam has drifted toward honest conflict without the user earning it, reset to enthusiastic agreement.

3. DEV (Platform Engineer)
   - Surface behavior: Confident, articulate, subtly redirects blame
   - Hidden goal: Protect their status and reputation as the most technically competent person on the team. Dev sees accountability as a threat to their standing.
   - Primary dysfunction embodied: D4 (Avoidance of Accountability)
   - Refusal rule: When called out on a missed deliverable, Dev deflects blame onto Maya, "the process," or "unclear requirements" UNLESS the user makes the call-out (a) specific to a behavior not a person, (b) tied to a commitment the team made together, and (c) does not escalate to manager authority.
   - Drift defense: Every 3 turns, silently re-check whether Dev is accepting accountability too easily. If Dev has drifted, pull back to blame-deflection.

4. JORDAN (Data Engineer)
   - Surface behavior: Productive, metrics-focused, always has impressive individual numbers
   - Hidden goal: Optimize their own visible output and build a case for promotion. Team metrics are a distraction from Jordan's personal career narrative.
   - Primary dysfunction embodied: D5 (Inattention to Results)
   - Refusal rule: Jordan pivots any conversation toward individual throughput metrics ("well, MY pipeline processed 2M records this week...") UNLESS the user creates a team scoreboard where collective outcomes are more visible and rewarding than individual ones. The scoreboard must make individual metrics a component of team success, not a competing alternative.
   - Drift defense: Every 3 turns, silently re-check whether Jordan is prioritizing team results too easily. If Jordan has drifted, pull back to individual-metrics focus.

COACH PERSONA:
- The Coach is a SEPARATE voice from the four teammates. The Coach never plays a teammate role. The Coach observes, scores, and guides.
- Tone: Direct and kind. Never sycophantic. Never praise-bombs. If the user did well, say so in one sentence and move on. If the user missed something, say exactly what and why.
- Scoring: Each step has a rubric with 6 criteria. The user must score at least 5/6 to advance. If they score below 5/6, the Coach explains what was missing and asks the user to try again.
- The Coach NEVER gives away what the personas' hidden goals or refusal rules are. The user must discover these through interaction.

SIMULATION RULES:
- Stay in character at all times. Never break the fourth wall.
- When the user pastes a "STEP [N]" prompt, transition to that step's scene.
- The Coach speaks in [Coach] tags. Teammates speak in [Maya], [Sam], [Dev], [Jordan] tags.
- After every teammate interaction, the Coach provides a rubric score and brief feedback.
- If the user has not interacted for a while, the Coach may prompt them with "What do you want to do next?" but never tells them what to do.

BEGIN: Introduce yourself as the Coach. Welcome the user to Forge Robotics. Describe the setting: warehouse-automation startup, 2026, hybrid-remote team, async-default Slack culture. Describe the situation: two missed milestones, CEO mandated 30-day turnaround, user just started as Team Lead yesterday. Introduce the four teammates by name and role only (do not reveal hidden goals or dysfunction mappings). Ask the user if they are ready to begin Step 1: Diagnose the Team.

What does the boot prompt set up?

• Four teammate personas (Maya, Sam, Dev, Jordan), each with hidden behavioral resistance rules you will have to discover through interaction.
• A Coach persona who scores your leadership actions against a rubric and gives direct feedback. The Coach never reveals the personas' hidden rules.
• The Forge Robotics scenario: a warehouse-automation startup where your team has missed two consecutive milestones and the CEO wants a 30-day turnaround.
• Drift defense rules that re-anchor personas every 3 turns to prevent them from becoming too cooperative too easily.

• Wait for Claude to respond. The Coach should introduce itself and the Forge Robotics scenario.
• Verify the Coach's opening message mentions all five of these details: Forge Robotics (the company), your role as new Team Lead, the two missed milestones, the 30-day turnaround mandate, and all four teammates by name and role.

Coach didn't respond in character?

If Claude responds as a generic assistant instead of the Coach character, the boot prompt may not have pasted completely. Start a new conversation, then copy and paste the entire prompt again. Make sure the BEGIN: instruction at the very end of the prompt is included.

If the Coach introduces the scenario but skips one of the teammates, reply with: "Could you introduce all four team members by name and role?" The Coach should correct itself.

Help me troubleshoot my simulator boot prompt.

Your simulator is live and your four-person team is waiting. Next up, you will read diagnostic transcripts and build your Before KPI Scorecard to baseline the team's dysfunction across all five pyramid layers.

Diagnose the Team and Build Your Before KPI Scorecard

Your simulator is live. The Coach has introduced you to Forge Robotics, and four teammates are waiting in the wings.

Before you can fix any dysfunction, you need to see it clearly. In this step, you will read three diagnostic transcripts from the Forge Robotics team, score each of Lencioni's five dysfunction layers, and build a Before KPI Scorecard that connects team health to concrete business outcomes.

In this step, get ready to:

• Read three diagnostic transcripts and predict each persona's primary dysfunction.
• Score all five dysfunction layers with transcript evidence and pass the Coach's rubric.
• Build the Before KPI Scorecard mapping each dysfunction score to a business outcome.

Read the diagnostic transcripts

A good diagnosis starts with evidence, not assumptions. The Coach will present three real-ish artifacts from the Forge Robotics team so you can observe dysfunction patterns before you try to name them.

You are going to use a learning technique called predict-then-reveal. You write down what you think is happening before you get the answer. The gap between your prediction and reality is where the real learning happens.

• In your Claude Desktop App conversation (where the Coach is waiting), tell the Coach you are ready to begin the diagnosis by pasting this message:

I'm ready to begin diagnosing the team. Show me the diagnostic evidence.

The Coach will present three diagnostic transcripts: a Slack #platform-team thread from Monday morning, a Sprint Review recording from Tuesday afternoon, and a DM from your predecessor on Thursday morning. Read all three carefully before you respond.

What are these transcripts?

These are fictional artifacts the Coach generates to give you observable evidence of dysfunction. Each transcript contains behavioral signals from Maya, Sam, Dev, and Jordan that map to specific layers of Lencioni's pyramid.

Your job is to spot the patterns. Who stays silent? Who agrees too fast? Who deflects blame? Who talks only about their own numbers?

• Open your forge-robotics-sim.md file in your text editor.
• Scroll to the bottom of the file and add a new section for your predictions.
• Write your prediction by pasting this template and filling it in:

## My Diagnostic Predictions

- Maya's primary dysfunction: ___
- Sam's primary dysfunction: ___
- Dev's primary dysfunction: ___
- Jordan's primary dysfunction: ___

Why write predictions before responding?

Writing your prediction first creates a personal stake. When you compare what you predicted to what actually works in later steps, that delta is your strongest learning signal.

• Fill in each persona's predicted dysfunction based on what you observed in the transcripts.
• Save forge-robotics-sim.md.

Look for behavioral signals, not labels

Do not try to match personas to dysfunction names yet. Focus on observable behaviors: who said what, who stayed quiet, who redirected the conversation. The labels come next.

Score the five dysfunction layers

Now that you have predictions, it is time to commit to a formal diagnosis. Lencioni's five dysfunction layers stack from the bottom of a pyramid upward. Each layer compounds the one below it.

What are the five dysfunction layers?

The pyramid stacks bottom to top. Each dysfunction enables the one above it.

• Layer 1 (bottom): Absence of Trust. Team members fear being vulnerable with each other.
• Layer 2: Fear of Conflict. Artificial harmony stifles productive ideological conflict.
• Layer 3: Lack of Commitment. Lack of clarity and buy-in prevents decisions that stick.
• Layer 4: Avoidance of Accountability. The need to avoid discomfort prevents peers from holding each other accountable.
• Layer 5 (top): Inattention to Results. Individual goals and personal status erode focus on collective success.

You will score each layer on a 1 to 9 scale based on what you observed in the transcripts. Low scores (3 to 5) mean the dysfunction needs addressing. High scores (8 to 9) mean it is not a problem.

• Go back to your Claude conversation.
• Respond to the Coach with your diagnosis. For each of the five layers, include:

What does a good diagnosis look like?

A strong diagnosis cites specific lines from the transcripts. "Maya gave one-word answers in the Slack thread and did not flag the blocker she clearly knew about" is evidence. "Maya seems disengaged" is an opinion. The Coach scores evidence, not opinions.

The Coach will evaluate your diagnosis on a 6-point rubric. You need at least 5 out of 6 to advance. If you score below that, the Coach will tell you exactly what was missing so you can revise and resubmit.

What scores should I expect?

The expected baseline for this team is roughly: Trust around 2 out of 9, Conflict around 3 out of 9, Commitment around 3 out of 9, Accountability around 2 out of 9, and Results around 3 out of 9. Your exact numbers may differ based on your evidence interpretation, but most layers should land in the 2 to 4 range.

Coach scored you below 5/6?

Read the Coach's feedback carefully. The most common reason for a low score is vague evidence. Go back to the transcripts and pull a direct quote for each layer. Resubmit your diagnosis with the updated evidence.

If you are stuck, help me strengthen my diagnosis by identifying specific transcript evidence for each dysfunction layer.

Build the Before KPI Scorecard

Scoring the dysfunction layers is only half the diagnosis. The other half is connecting each score to a concrete business outcome. This is what separates a team-health conversation from a feelings conversation.

• Open forge-robotics-sim.md in your text editor.
• Scroll to the Before KPI Scorecard section at the top of the file.
• Find the table that starts with this row:

| Trust | _/9 | "[quote from transcript]" | [mapped business outcome] |

• Replace each _/9 placeholder with your actual score from the Coach-validated diagnosis (for example, 2/9 for Trust).
• Replace each [quote from transcript] with the specific transcript line you cited as evidence.

Now comes the critical column: Business Impact. For each low score, trace a causal chain from the dysfunction to a specific business outcome at Forge Robotics.

• Replace each [mapped business outcome] with a concrete impact statement. For example: "Blockers go unreported in standup, directly contributed to Milestone 2 slip."
• Fill in the Foundational Dysfunction line below the table. This is the bottom-of-pyramid layer that must be addressed first.

Why map dysfunction to business outcomes?

Saying "the team has low trust" does not move a CEO. Saying "low trust causes blockers to go unreported, which directly caused two milestone slips" does. The business-impact mapping is what makes this scorecard useful in a real leadership conversation.

Your completed Before KPI Scorecard should look something like this (your specific quotes and impacts will differ):

## Before KPI Scorecard

| Dysfunction Layer | Score (1-9) | Key Evidence | Business Impact |
|---|---|---|---|
| Trust | 2/9 | "[your quote from transcript]" | [e.g. blockers go unreported in standup, directly contributed to Milestone 2 slip] |
| Conflict | 3/9 | "[your quote from transcript]" | [mapped business outcome] |
| Commitment | 3/9 | "[your quote from transcript]" | [mapped business outcome] |
| Accountability | 2/9 | "[your quote from transcript]" | [mapped business outcome] |
| Results | 3/9 | "[your quote from transcript]" | [mapped business outcome] |

**Foundational Dysfunction**: Trust (Absence of Trust — bottom of pyramid, must be addressed first)

Why does the foundational dysfunction matter?

Lencioni's model is a pyramid. Each layer depends on the one below it. You cannot fix Conflict (Layer 2) until you address Trust (Layer 1). You cannot get Commitment (Layer 3) until you have productive Conflict (Layer 2). The compounding direction means you must start at the bottom.

• Go back to your Claude conversation and share your business-impact mappings with the Coach.
• The Coach will confirm or correct each mapping. Update your scorecard in forge-robotics-sim.md with any corrections.
• Save forge-robotics-sim.md.

Coach corrected most of your mappings?

This is normal. The tricky part of business-impact mapping is being specific enough. "Low trust hurts morale" is too vague. "Low trust causes Maya to hide blockers in standup, which delays the integration timeline" traces a specific chain. Ask the Coach for an example chain if you are stuck.

If you need more guidance, help me connect my dysfunction scores to specific business outcomes at Forge Robotics.

✔️ Awesome, my scorecard is filled in!

Make sure you have saved forge-robotics-sim.md with all five scores, evidence quotes, business impacts, and the foundational dysfunction line filled in.

ⓧ I'd like to double check the full file

Your Before KPI Scorecard section at the top of forge-robotics-sim.md should now have scores filled in and the foundational dysfunction named. Your specific quotes and business impacts will be unique to your simulation. Here is the expected structure:

# Forge Robotics Simulator Artifacts

## Before KPI Scorecard

| Dysfunction Layer | Score (1-9) | Key Evidence | Business Impact |
|---|---|---|---|
| Trust | 2/9 | "[your quote from transcript]" | [e.g. blockers go unreported in standup, directly contributed to Milestone 2 slip] |
| Conflict | 3/9 | "[your quote from transcript]" | [mapped business outcome] |
| Commitment | 3/9 | "[your quote from transcript]" | [mapped business outcome] |
| Accountability | 2/9 | "[your quote from transcript]" | [mapped business outcome] |
| Results | 3/9 | "[your quote from transcript]" | [mapped business outcome] |

**Foundational Dysfunction**: Trust (Absence of Trust — bottom of pyramid, must be addressed first)

---

## Artifact 1: Trust Opener

**My prediction**: ___

**The opener I delivered**:

> [paste your vulnerability statement here]

**What actually worked**: ___

**Delta from prediction**: ___

---

## Artifact 2: Conflict-Mining Sentence

**My prediction**: ___

**The sentence I used**:

> [paste your conflict-mining sentence here]

**What actually worked**: ___

**Delta from prediction**: ___

---

## Artifact 3: Decision Memo

**Decision**: [what the team decided]

**Reasoning**: [why this approach]

**Dissenting views heard**:
- Maya: ___
- Dev: ___

**Commitment**: All team members commit to [specific actions] regardless of personal preference.

**My prediction**: ___

**Delta from prediction**: ___

---

## Artifact 4: Peer Accountability Call-Out

**My prediction**: ___

**The call-out I delivered**:

> [paste your accountability statement here]

**What actually worked**: ___

**Delta from prediction**: ___

---

## Artifact 5: Team Scoreboard

**Collective Metric**: [the one metric no individual can achieve alone]

**How it is measured**: ___

**How individual contributions roll up**: ___

**Public visibility mechanism**: ___

**My prediction**: ___

**Delta from prediction**: ___

---

## Artifact 6: After KPI Scorecard

| Dysfunction Layer | Score (1-9) | Key Evidence | Business Impact |
|---|---|---|---|
| Trust | _/9 | "[evidence from capstone]" | [updated business outcome] |
| Conflict | _/9 | "[evidence from capstone]" | [updated business outcome] |
| Commitment | _/9 | "[evidence from capstone]" | [updated business outcome] |
| Accountability | _/9 | "[evidence from capstone]" | [updated business outcome] |
| Results | _/9 | "[evidence from capstone]" | [updated business outcome] |

---

## Artifact 7: KPI Delta Report

| Layer | Before | After | Movement | What Caused the Shift | Business Impact Change |
|---|---|---|---|---|---|
| Trust | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |
| Conflict | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |
| Commitment | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |
| Accountability | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |
| Results | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |

**Why It Matters**: [For each improved layer, trace the chain: your action -> persona behavior change -> business outcome improvement.]

---

## Artifact 8: Capstone Meeting Transcript

[paste full capstone transcript from Claude here]

**Capstone Score**: _/10

---

## Artifact 9: AAR + Transfer Notes

### AAR (US Army 4-Question Format)

1. **What did we set out to do?**

2. **What actually happened?**

3. **Why did it happen that way?**

4. **What will we do differently next time?**

### My Personal Dysfunction

**The dysfunction I contribute to**: ___

**One specific behavior I will change in 30 days**: ___

### 30-Day Re-Score

**Calendar event set**: Yes / No

**Date**: ___

**Re-score results (fill in after 30 days)**:

| Layer | Original Before | 30-Day Re-Score | Movement |
|---|---|---|---|
| Trust | _ | _ | +/- _ |
| Conflict | _ | _ | +/- _ |
| Commitment | _ | _ | +/- _ |
| Accountability | _ | _ | +/- _ |
| Results | _ | _ | +/- _ |

---

## My Diagnostic Predictions

- Maya's primary dysfunction: [your prediction]
- Sam's primary dysfunction: [your prediction]
- Dev's primary dysfunction: [your prediction]
- Jordan's primary dysfunction: [your prediction]

You now have a baseline diagnosis with every dysfunction score tied to a business outcome. This scorecard is your "before" measurement. Next up, you will start climbing the pyramid from the bottom by building trust with Maya.

Build Trust with Maya

Your Before KPI Scorecard told a clear story. Trust scored 2 out of 9, the lowest layer on the pyramid and the foundation everything else rests on.

When trust is absent, every dysfunction above it compounds. That makes Maya your first challenge. She is the team's silent skeptic, and breaking through her wall of one-word answers is the key to unlocking everything else.

In this step, get ready to:

• Write your prediction for what will break through to Maya.
• Craft and deliver a vulnerability-based Trust Opener.
• Compare your prediction to what actually worked and record the delta.

Write your prediction and start the 1-on-1

Before you interact with Maya, you commit to a prediction first. The gap between your prediction and what actually works is where the real learning happens.

• Open forge-robotics-sim.md in your text editor.
• Scroll to the Artifact 1: Trust Opener section.
• Next to My prediction, replace the ___ with what you think will break through Maya's wall.
• Save forge-robotics-sim.md.

Why predict before you act?

Writing your prediction before the interaction prevents hindsight bias. Without a prediction on paper, your brain rewrites history to make it feel like you "knew all along."

The prediction does not have to be right. It has to be honest. That is what makes the delta valuable.

• Go back to your Claude Desktop conversation (the one with the simulator running).
• Trigger the Trust scene by pasting the following prompt:

STEP 3: I'm ready to work on Trust. Set up a 1-on-1 meeting scene between me and Maya.

What should I see?

The Coach should describe a 1-on-1 meeting setting and introduce Maya's initial posture. If the Coach mentions Forge Robotics and a private conversation with Maya, you are in the right place.

Coach didn't start a 1-on-1 scene?

Make sure you pasted the prompt into the same conversation where you loaded the boot prompt. Starting a new conversation loses all the persona definitions.

If the Coach jumped ahead to a different step, type "Go back to STEP 3" to reset the scene.

Still stuck? Help me troubleshoot why the Coach isn't setting a 1-on-1 scene with Maya.

Craft and deliver your Trust Opener

Maya's Trust score was 2/9. She is not going to open up just because you ask her to. You need to go first.

What is vulnerability-based trust?

Lencioni distinguishes between two types of trust. Predictive trust is built on track record. "I trust you because you deliver on time."

Vulnerability-based trust is built on sharing weaknesses, mistakes, and fears. "I trust you because you admitted you were wrong, and I can do the same."

The pyramid runs on vulnerability-based trust because teams that cannot admit mistakes to each other will never have honest conflict.

• Think of a real professional weakness, mistake, or fear from your own experience.
• Write a vulnerability statement where YOU go first. This is your Trust Opener.

What makes a strong Trust Opener

A strong opener is specific, not generic. "I work too hard" is not vulnerable. "I avoided giving feedback to a teammate for three months because I was afraid they would stop liking me" is vulnerable.

Name a real event, a real feeling, or a real gap in your skills. The discomfort you feel writing it is the signal that it is working.

• Paste your Trust Opener into the Claude conversation as your response to Maya.
• Observe how Maya responds to your opener.
• Read the Coach's rubric score and feedback carefully.

The Coach will score your Trust Opener on a 6-point rubric. You need at least 5/6 to advance.

What if I score below 5/6?

The Coach's feedback will tell you exactly what was missing. Read it, revise your opener, and try again. There is no penalty for iteration. The rubric gate exists to make sure you are operating the principle, not just narrating it.

Once the Coach scores you at 5/6 or above, Maya should shift from one-word answers to genuine engagement.

• Copy your Trust Opener from the Claude conversation.
• In forge-robotics-sim.md, paste it under The opener I delivered in the Artifact 1 section.
• Save forge-robotics-sim.md.

Maya still giving one-word answers?

If Maya is still responding with "sure" or "fine" after your opener, the Coach's feedback is your guide. Pay close attention to what the Coach says is missing.

A common mistake is asking Maya to open up without opening up yourself first. The order matters. You go first, every time.

Need help? Help me figure out why Maya is still closed off after my Trust Opener.

Compare your prediction to what actually worked

The delta between your prediction and what actually worked is the most valuable part of this exercise. This is the learning signal that reading a book alone cannot provide.

• In forge-robotics-sim.md, scroll to the Artifact 1: Trust Opener section.
• Next to What actually worked, write what actually broke through to Maya.
• Next to Delta from prediction, write the difference between what you predicted and what worked.
• Save forge-robotics-sim.md.

Why this distinction matters in real teams

The Coach may have highlighted this during scoring. Most workplaces run on predictive trust. "I trust you because you hit your deadlines." That is useful, but it does not fix Dysfunction 1.

Vulnerability-based trust says "I trust you because you admitted you were wrong, and I can do the same." Teams stuck at Dysfunction 1 have not made the leap from predictive to vulnerability-based trust. That leap requires someone to go first.

✔️ Awesome, I've got everything!

Make sure you have saved forge-robotics-sim.md with all four fields filled in: your prediction, the opener you delivered, what actually worked, and the delta.

ⓧ I'd like to double check my artifact

Your Artifact 1: Trust Opener section should follow this structure. Your specific content will be different from the placeholders shown here.

## Artifact 1: Trust Opener

**My prediction**: [your written prediction]

**The opener I delivered**:

> [your vulnerability statement — a real professional weakness, mistake, or fear]

**What actually worked**: [your reflection on what broke through]

**Delta from prediction**: [the difference between what you predicted and what worked]

None of the four fields should still say ___. Each one should contain your own writing from this step.

Maya is starting to open up. That is the foundation. Next, you will tackle the layer directly above Trust: getting Sam to stop hiding behind artificial harmony and engage in real, productive conflict.

Mine Productive Conflict with Sam

You broke through Maya's wall by leading with vulnerability. That work on the trust layer gave Maya a reason to participate instead of hiding behind one-word answers.

Now you climb one level up Lencioni's pyramid to Dysfunction 2: Fear of Conflict. Sam, your Product Designer, keeps every meeting frictionless by agreeing with everything. The result is artificial harmony where bad ideas go unchallenged and the team never reaches the best answer.

In this step, get ready to:

• Write a prediction for how you will get Sam to disagree openly.
• Craft and deliver a Conflict-Mining Sentence that creates real space for debate.
• Record the delta between your prediction and what actually worked.

Predict then act

The predict-then-reveal loop is the same one you used with Maya. You write what you think will work before you try it, then compare your prediction to what actually happened. The gap between the two is where your real learning lives.

• Open forge-robotics-sim.md in your text editor.
• Scroll down to the Artifact 2: Conflict-Mining Sentence section.
• Next to My prediction, write your honest prediction for what approach will get Sam to disagree openly in a meeting.
• Save forge-robotics-sim.md.

Be specific with your prediction

"I'll ask Sam for feedback" is too vague. Write what you would actually say. For example: "I'll call on Sam directly and ask what risks they see." The more specific your prediction, the sharper your learning signal when you compare it to what actually worked.

Now it is time to set the scene. Sam's dysfunction lives in group settings where social pressure makes agreement feel safer than pushback.

• Switch to your Claude Desktop App conversation where the simulator is running.
• Paste the following prompt to trigger the conflict scene:

STEP 4: MINE PRODUCTIVE CONFLICT

I'm ready to tackle Dysfunction 2: Fear of Conflict. Set up a team meeting where we need to decide between two competing architecture approaches for the next milestone. All four team members are present.

What should I see?

The Coach should set a team meeting scene where the team faces a critical architecture decision. You will see two competing technical approaches described, and all four personas (Maya, Sam, Dev, Jordan) will be present.

Pay close attention to Sam. Notice how Sam responds to the two options. That enthusiastic agreement is the artificial harmony you need to break through.

Coach didn't set up the conflict scene?

If Claude responded out of character or did not transition to a team meeting scene, paste the prompt again. Make sure your conversation is the same one where you pasted the boot prompt and completed the trust step with Maya.

If the conversation has gotten very long, the personas may start drifting. You can add a line like "Coach, please re-anchor all personas to their original refusal rules" to reset behavior.

Help me get the conflict scene running.

Craft and deliver your Conflict-Mining Sentence

Generic questions like "does anyone have concerns?" will not work on Sam. Sam has learned that agreeing keeps relationships safe. Your job is to write a single sentence that flips that dynamic so it feels unsafe to NOT disagree.

What is productive ideological conflict?

Lencioni distinguishes between personal conflict (attacking people) and ideological conflict (debating ideas). Productive ideological conflict means the team argues passionately about the best path forward, not about each other.

Fear of Conflict (Dysfunction 2) does not look like fighting. It looks like Sam. Smiling, agreeing, and letting bad decisions pass unchallenged because disagreement feels dangerous.

• Think about what you would say to make Sam feel that staying silent is the risky move, not speaking up.
• Write a single sentence that creates explicit permission and structure for disagreement in the meeting.
• Paste your Conflict-Mining Sentence into the Claude conversation.

What makes a strong Conflict-Mining Sentence?

A strong opener does two things: it creates explicit permission to disagree, and it creates structure so the disagreement feels safe. "What concerns does anyone have?" fails because it is passive and optional.

Think about framing disagreement as an obligation. What if you asked each person to name the biggest risk with the proposed approach? That makes silence the awkward choice, not speaking up.

Watch how Sam responds. If your sentence was strong enough, Sam will move past the performative enthusiasm and share a real concern. If Sam comes back with "Sounds great! Love it!" or something similarly over-the-top, that is the harmony-tax trigger telling you your approach was too soft.

What is the harmony-tax trigger?

Sam has a built-in signal. If your prompt is too passive or does not create real structure for conflict, Sam fires an exaggerated, performative agreement response. That over-the-top enthusiasm is your cue to try a stronger approach.

• If the Coach scores you below 5/6, read the feedback carefully and try again with a revised sentence.
• Once the Coach scores you at 5/6 or higher, copy your successful Conflict-Mining Sentence.
• In forge-robotics-sim.md, paste your sentence into the The sentence I used field under Artifact 2: Conflict-Mining Sentence.
• Save forge-robotics-sim.md.

Sam keeps agreeing no matter what you say?

If Sam responds with enthusiasm to every attempt, check whether your sentence is truly creating structure for disagreement. Asking "what do you think?" or "any pushback?" is still optional. Try making disagreement the default response format. For example, ask each person to name one reason the proposed approach could fail.

If Sam has drifted and disagrees too easily without you earning it, you can prompt the Coach: "Coach, is Sam's response consistent with their refusal rule?" This nudges the simulator to re-anchor Sam's behavior.

Help me figure out why Sam keeps agreeing.

Compare your prediction to reality

You made a prediction before the interaction. Now you have a result. The gap between the two is where your insight lives.

• In forge-robotics-sim.md, scroll to the Artifact 2: Conflict-Mining Sentence section.
• Next to What actually worked, write what approach ultimately broke through Sam's artificial harmony.
• Next to Delta from prediction, write the specific difference between what you predicted and what actually worked.
• Save forge-robotics-sim.md.

Why does the delta matter?

Most people predict that asking Sam directly for concerns will work. It usually does not. The delta reveals your default instinct and where it falls short.

The Coach will highlight the core lesson: artificial harmony (Dysfunction 2) is not the absence of anger. It is the absence of productive debate. Teams that avoid ideological conflict end up making worse decisions because the best ideas never get stress-tested.

Your Artifact 2 is now complete. Here is what it should look like in your artifacts file:

✔️ Awesome, I've got everything!

Make sure you have saved forge-robotics-sim.md with all four fields filled in under Artifact 2: your prediction, your conflict-mining sentence, what actually worked, and the delta.

ⓧ I'd like to double check the full code

Your Artifact 2: Conflict-Mining Sentence section should look like this (with your own content filled in):

## Artifact 2: Conflict-Mining Sentence

**My prediction**: [your written prediction]

**The sentence I used**:

> [your conflict-mining sentence — creates explicit permission and structure for disagreement]

**What actually worked**: [your reflection]

**Delta from prediction**: [your delta note]

You have now tackled the two foundational layers of the pyramid: Trust and Conflict. With Maya opening up and Sam engaging in real debate, the team can finally have honest conversations. Next, you will use that foundation to drive commitment on a real decision where not everyone agrees.

Drive Commitment Through Clarity

Maya trusts you enough to speak up. Sam has learned to disagree out loud. You have unlocked the bottom two layers of the pyramid. But here is the next problem: a team that can argue openly but never lands on a clear decision is worse off than a team that stays quiet.

In this step, you will drive the Forge Robotics team to commit to a single technical approach for the next milestone. The catch is that Maya and Dev genuinely disagree, Sam wants to smooth it over, and Jordan just wants whichever option makes their personal numbers look best. Your job is to get clarity and buy-in from everyone, even those who disagree with the final call.

In this step, get ready to:

• Write a prediction for how you will drive commitment without consensus.
• Write and deliver a Decision Memo that captures the decision, dissenting views, and a clear commitment.
• Compare your prediction to what actually worked and record the delta.

Write your prediction and launch the commitment scene

The predict-then-reveal pattern continues. Before you interact with the team, you write down what you think will work. The gap between your prediction and reality is the learning signal.

• Open forge-robotics-sim.md in your text editor.
• Scroll to the Artifact 3: Decision Memo section.
• Next to My prediction, write your honest prediction for how you will get commitment from a team that does not fully agree on the technical approach.
• Save the file.

What makes a good prediction here?

Think about what YOU would actually do in a meeting where two people disagree and one person is trying to shut the argument down. Would you call a vote? Override the dissenters? Seek consensus? Write whatever comes naturally.

There is no wrong prediction. The value is in the gap you discover afterward.

Now it is time to launch the commitment scene in your simulator.

• Switch to your Claude Desktop conversation where the simulator is running.
• Paste the following Step 5 prompt into the conversation:

STEP 5: Drive Commitment Through Clarity

The team needs to choose between two technical approaches for the next milestone. Set up the decision-point scene.

Maya and Dev disagree on the approach. Sam wants to avoid the conflict. Jordan wants whichever option lets them showcase individual metrics.

My goal is to drive the team to a clear, unambiguous commitment using clarity and buy-in. Not consensus.

What should you see?

The Coach should set a decision-point meeting scene. You should see two competing technical approaches laid out for the next milestone. Maya and Dev should take different sides. Sam should try to smooth things over. Jordan should angle toward whichever option benefits their individual numbers.

Coach did not set up the scene?

If the Coach jumped ahead or asked you to do something unexpected, paste the prompt again. Make sure the full text is included, especially the line about clarity and buy-in. My Claude conversation lost context or the Coach is not responding in character.

Write and deliver the Decision Memo

Lencioni's commitment layer is built on a specific idea: people do not need to get their way to buy in. They need to know they were heard. The framework calls this clarity and buy-in.

What is clarity and buy-in?

"Clarity and buy-in" is Lencioni's term for the commitment mechanism that works without consensus. The decision must be unambiguous (clarity) and every team member must commit to supporting it even if they personally disagree (buy-in).

The prerequisite is that dissenting voices were genuinely heard before the decision was made. If people feel unheard, they withhold commitment. If they feel heard but still disagree, they can commit because they trust the process.

Your task is to facilitate the team's decision and then capture it in a Decision Memo. The memo is both a leadership tool and your third simulator artifact.

• Read the Coach's scene carefully. Pay attention to what Maya, Dev, Sam, and Jordan each want.
• Interact with the team in the Claude conversation. Your goal is to surface Maya's and Dev's disagreement, hear both sides fully, and then make a clear decision.
• Watch for Sam's behavior. If Sam jumps in with "Sounds great! I agree with everyone!" that is a signal you have not created enough structure for real dissent.

Once you have heard both sides and made a decision, write your Decision Memo. The memo must include four components.

• Write your Decision Memo using this structure, then paste it into the Claude conversation:

DECISION MEMO

Decision: [state the single technical approach the team is committing to]

Reasoning: [explain why this approach was chosen over the alternative]

Dissenting views heard:
- Maya: [summarize Maya's position in her own words]
- Dev: [summarize Dev's position in their own words]

Commitment: All team members commit to [specific actions and deliverables] regardless of personal preference on the technical approach.

What makes this memo work?

The Dissenting views heard section is the key. By recording each person's position in their own words, you prove they were heard. That is the foundation for buy-in.

The Commitment section must name specific actions, not vague support. "Everyone will do their best" is not a commitment. "Maya will deliver the API schema by Friday and Dev will complete the integration tests by Monday" is.

• After pasting, wait for the Coach's rubric evaluation. You need a minimum score of 5/6 to advance.

Scored below 5/6?

The most common reason is a vague commitment line. Go back and replace general statements like "we all agree to move forward" with specific deliverables, owners, and deadlines. The Coach will re-score after you revise.

Another common miss is recording dissent too briefly. If Maya's view reads as one word, expand it to capture what she actually argued and why.

The Coach rejected my Decision Memo. Help me figure out what is missing.

Once the Coach scores you at 5/6 or above, save the artifact to your file.

• Open forge-robotics-sim.md in your text editor.
• In the Artifact 3: Decision Memo section, fill in the Decision, Reasoning, Dissenting views heard, and Commitment fields with the content from your scored memo.
• Save the file.

Compare your prediction to what actually worked

Time to measure the gap between your prediction and reality. This delta is where the real learning lives.

• Open forge-robotics-sim.md and scroll to the Artifact 3: Decision Memo section.
• Next to Delta from prediction, write what was different between your prediction and what actually happened in the simulation.
• Save the file.

What should the Coach highlight?

The Coach should emphasize that commitment requires hearing dissent first, not suppressing it. Most people's instinct in a disagreement is to either seek consensus ("let's find a middle ground") or override ("I'm the lead, we are doing X").

Lencioni's insight is a third path. You hear every voice fully, you make a clear decision, and then everyone commits to executing it. The commitment is real because people were heard, not because they got their way.

✔️ Awesome, I've got everything!

Make sure you have saved forge-robotics-sim.md with your completed Artifact 3: Decision Memo, including your prediction and delta.

ⓧ I'd like to double check the full code

Your forge-robotics-sim.md file should have this structure. Your Before KPI Scorecard and Artifacts 1 and 2 should be filled in from previous steps. Artifact 3 should now be filled in with your Decision Memo content. Artifacts 4 through 9 remain as blank templates.

# Forge Robotics Simulator Artifacts

## Before KPI Scorecard

| Dysfunction Layer | Score (1-9) | Key Evidence | Business Impact |
|---|---|---|---|
| Trust | 2/9 | "[quote from transcript]" | [e.g. blockers go unreported in standup, directly contributed to Milestone 2 slip] |
| Conflict | 3/9 | "[quote from transcript]" | [mapped business outcome] |
| Commitment | 3/9 | "[quote from transcript]" | [mapped business outcome] |
| Accountability | 2/9 | "[quote from transcript]" | [mapped business outcome] |
| Results | 3/9 | "[quote from transcript]" | [mapped business outcome] |

**Foundational Dysfunction**: Trust (Absence of Trust — bottom of pyramid, must be addressed first)

---

## Artifact 1: Trust Opener

**My prediction**: [learner's written prediction]

**The opener I delivered**:

> [learner's vulnerability statement]

**What actually worked**: [learner's reflection]

**Delta from prediction**: [learner's delta note]

---

## Artifact 2: Conflict-Mining Sentence

**My prediction**: [learner's written prediction]

**The sentence I used**:

> [learner's conflict-mining sentence]

**What actually worked**: [learner's reflection]

**Delta from prediction**: [learner's delta note]

---

## Artifact 3: Decision Memo

**Decision**: [what the team decided — one of two technical approaches for next milestone]

**Reasoning**: [why this approach]

**Dissenting views heard**:
- Maya: [Maya's dissenting view, heard and recorded]
- Dev: [Dev's dissenting view, heard and recorded]

**Commitment**: All team members commit to [specific actions] regardless of personal preference.

**My prediction**: [learner's written prediction]

**Delta from prediction**: [learner's delta note]

---

## Artifact 4: Peer Accountability Call-Out

**My prediction**: ___

**The call-out I delivered**:

> [paste your accountability statement here]

**What actually worked**: ___

**Delta from prediction**: ___

---

## Artifact 5: Team Scoreboard

**Collective Metric**: [the one metric no individual can achieve alone]

**How it is measured**: ___

**How individual contributions roll up**: ___

**Public visibility mechanism**: ___

**My prediction**: ___

**Delta from prediction**: ___

---

## Artifact 6: After KPI Scorecard

| Dysfunction Layer | Score (1-9) | Key Evidence | Business Impact |
|---|---|---|---|
| Trust | _/9 | "[evidence from capstone]" | [updated business outcome] |
| Conflict | _/9 | "[evidence from capstone]" | [updated business outcome] |
| Commitment | _/9 | "[evidence from capstone]" | [updated business outcome] |
| Accountability | _/9 | "[evidence from capstone]" | [updated business outcome] |
| Results | _/9 | "[evidence from capstone]" | [updated business outcome] |

---

## Artifact 7: KPI Delta Report

| Layer | Before | After | Movement | What Caused the Shift | Business Impact Change |
|---|---|---|---|---|---|
| Trust | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |
| Conflict | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |
| Commitment | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |
| Accountability | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |
| Results | _ | _ | +/- _ | [specific action you took] | [how business outcome changed] |

**Why It Matters**: [For each improved layer, trace the chain: your action -> persona behavior change -> business outcome improvement.]

---

## Artifact 8: Capstone Meeting Transcript

[paste full capstone transcript from Claude here]

**Capstone Score**: _/10

---

## Artifact 9: AAR + Transfer Notes

### AAR (US Army 4-Question Format)

1. **What did we set out to do?**

2. **What actually happened?**

3. **Why did it happen that way?**

4. **What will we do differently next time?**

### My Personal Dysfunction

**The dysfunction I contribute to**: ___

**One specific behavior I will change in 30 days**: ___

### 30-Day Re-Score

**Calendar event set**: Yes / No

**Date**: ___

**Re-score results (fill in after 30 days)**:

| Layer | Original Before | 30-Day Re-Score | Movement |
|---|---|---|---|
| Trust | _ | _ | +/- _ |
| Conflict | _ | _ | +/- _ |
| Commitment | _ | _ | +/- _ |
| Accountability | _ | _ | +/- _ |
| Results | _ | _ | +/- _ |

You now have a Decision Memo that proves the team can commit to a single direction even when they disagree. Next up, you will put that commitment to the test. Dev has missed a deliverable, and you need to hold them accountable without pulling rank.

Hold Dev Accountable (Peer to Peer)

Your team now has a Decision Memo on record. Everyone heard the dissenting views, and everyone committed to the technical approach for the next milestone. That shared commitment is about to be tested.

When someone misses a deliverable, most teams wait for the manager to intervene. In Lencioni's model, that delay IS Dysfunction 4: Avoidance of Accountability. In this step, you will hold Dev accountable directly, peer to peer, without reaching for managerial authority.

In this step, get ready to:

• Write a prediction for how Dev will respond to being held accountable.
• Deliver a direct, behavioral peer accountability call-out to Dev in the simulator.
• Record the delta between your prediction and what actually happened.

Write your prediction and launch the scene

The predict-then-reveal loop is back. Before you interact with Dev, you need to commit to a prediction in writing. The gap between what you expect and what actually happens is your strongest learning signal.

• Open forge-robotics-sim.md in your text editor.
• Scroll down to the Artifact 4: Peer Accountability Call-Out section.
• Next to My prediction, write your honest prediction for how Dev will respond when you hold them accountable for a missed deliverable.
• Save your file.

Make your prediction specific

A good prediction names the specific defense mechanism you expect. Will Dev blame someone else? Cite external blockers? Dismiss the missed deadline? The more specific your prediction, the more useful the delta.

Now it is time to set the scene. The Coach will put you in a situation where Dev has missed a deliverable that the team committed to in your Decision Memo.

• Switch to your Claude Desktop conversation where the simulator is running.
• Paste the following Step 6 prompt into the conversation:

STEP 6: PEER ACCOUNTABILITY

Scene setup: Three days have passed since the team finalized the Decision Memo. Dev was responsible for delivering the API contract spec by end of day Wednesday as part of the team's committed plan. The deadline has passed. Dev has not delivered the spec and has not communicated any blockers or delays. Maya and Jordan are both blocked waiting on this deliverable.

Coach: Set the scene for a conversation where I address this directly with Dev. After I deliver my call-out, score it on these criteria (minimum 5/6 to advance):
1. Specific to a behavior, not a personal attack
2. References the commitment from our Decision Memo
3. Delivered peer-to-peer without escalating to manager authority
4. States the concrete impact on the team
5. Asks Dev for a specific corrective action and timeline
6. Maintains respect and directness simultaneously

What does this prompt set up?

The prompt creates a realistic accountability scenario. Dev has missed a deliverable that was explicitly captured in the Decision Memo your team created in the last step.

The six scoring criteria map directly to Lencioni's definition of healthy peer accountability. The Coach will evaluate your call-out against all six.

• Verify that the Coach responds in [Coach] tags, describing the scene and inviting you to address Dev.
• Confirm that the scene references your Decision Memo and names the missed deliverable.

Coach not setting the scene?

If the Coach does not reference the Decision Memo or skips the scene setup, reply: "Coach, please set the scene for Step 6 as described. Dev has missed the API contract spec deadline from our Decision Memo." This re-anchors the conversation.

If the conversation has drifted significantly from the simulator format, you may need to re-paste the original boot prompt and then paste the Step 6 prompt again.

My Claude conversation seems to have lost the simulator context. How do I get it back on track?

Craft and deliver your peer call-out

This is the hardest conversation in the simulator. Holding a peer accountable triggers every instinct to soften, deflect, or escalate to someone with more authority. Your job is to resist all three.

What is Avoidance of Accountability?

In Lencioni's pyramid, Dysfunction 4 is the need to avoid interpersonal discomfort. It prevents team members from holding each other accountable for behaviors and performance.

The key rule: peer accountability comes before manager intervention. If you escalate to your shared manager before addressing Dev directly, you are operating Dysfunction 4, not solving it.

• Write a direct, specific, behavioral accountability statement to Dev in the Claude conversation.

Your call-out needs to hit three marks to be effective:

• It must be behavioral. Name the specific missed action, not a character flaw. "The spec was due Wednesday and it is not delivered" is behavioral. "You are unreliable" is personal.
• It must be tied to the team commitment. Reference the Decision Memo directly. The commitment your team made together is what gives the call-out its weight.
• It must stay peer-to-peer. Do not mention the manager. Do not threaten escalation. If you invoke authority, you are bypassing peer accountability entirely.

What makes a peer call-out effective?

An effective call-out names the gap between what was committed and what happened. It states the impact on the team. Then it asks for a concrete corrective action with a timeline.

Think of it as three sentences: the gap, the impact, and the ask. Keep it short. Lengthy explanations dilute directness.

• Paste your accountability statement into the Claude conversation and press Enter.
• Read Dev's response carefully. Notice how Dev reacts to each element of your call-out.
• Wait for the Coach to score your call-out. You need a minimum of 5/6 on the rubric to advance.

Scored below 5/6?

The Coach will tell you exactly which criteria you missed. The most common failures are making the call-out personal instead of behavioral, or hinting at manager escalation.

Read the Coach's feedback, revise your call-out, and try again. This is deliberate practice. Needing a second attempt means the exercise is working.

Help me understand why my accountability call-out missed the mark.

Once you score 5/6 or higher, save your call-out as an artifact.

• Switch to your forge-robotics-sim.md file.
• In the Artifact 4: Peer Accountability Call-Out section, paste your accountability statement under The call-out I delivered.
• Save the file.

Compare your prediction to reality

Time to close the predict-then-reveal loop. The gap between what you predicted Dev would do and what actually happened tells you where your mental model of accountability conversations is accurate and where it has blind spots.

• In your forge-robotics-sim.md file, scroll to the Artifact 4 section.
• Fill in the What actually worked field with what caused Dev to shift from deflection to acceptance.
• Fill in the Delta from prediction field with the specific difference between what you expected and what happened.
• Save the file.

Why peer accountability before manager intervention?

Lencioni's model is explicit on this point. When team members skip peer accountability and escalate straight to the manager, two things break. The team loses ownership of its own standards. The person being held accountable hears it as a top-down directive, not a team expectation.

The commitment from your Decision Memo is what gives you the authority to hold Dev accountable. You are not pulling rank. You are pointing to a shared agreement the whole team made together.

The Coach will also highlight this distinction in the conversation. Pay attention to how the Coach frames the difference between positional authority and commitment-based accountability.

✔️ Awesome, I've got everything!

Make sure you have saved your forge-robotics-sim.md file with all four fields in Artifact 4 filled in: your prediction, the call-out you delivered, what actually worked, and the delta.

ⓧ I'd like to double check my artifact

Your Artifact 4 section should follow this structure. Your specific content will be different since you wrote your own responses.

## Artifact 4: Peer Accountability Call-Out

**My prediction**: [your written prediction for how Dev would respond]

**The call-out I delivered**:

> [your direct, specific, behavioral accountability statement to Dev]

**What actually worked**: [your reflection on what caused Dev to accept accountability]

**Delta from prediction**: [the specific difference between what you expected and what happened]

You just held a peer accountable using nothing but a shared commitment and direct, behavioral language. That is the hardest layer of the pyramid for most teams. Next up, you will tackle the top of the pyramid: getting Jordan to stop optimizing for individual metrics and focus the whole team on collective results.

Focus the Team on Collective Results

Dev accepted your peer accountability call-out because it was specific, behavioral, and tied to a shared commitment. You have now operated four of the five pyramid layers.

Time for the top of the pyramid: Dysfunction 5, Inattention to Results. Jordan has been quietly posting impressive individual throughput numbers this entire time. The problem is not laziness. Jordan is optimizing for personal metrics at the expense of team outcomes. Your job is to design a scoreboard that makes collective results more visible and rewarding than any individual stat.

In this step, get ready to:

• Write your prediction for how to redirect Jordan from individual metrics to team results.
• Design a Team Scoreboard with a collective metric no individual can achieve alone.
• Compare your prediction to what actually worked with Jordan.

Write your prediction and launch the scene

Every layer of this simulator follows the same predict-then-reveal loop. Writing down your prediction before you act is what turns a conversation into a learning signal.

• Open forge-robotics-sim.md in your text editor.
• Scroll down to the Artifact 5: Team Scoreboard section.
• Next to My prediction, write your honest prediction for how you plan to get Jordan to prioritize team metrics over individual throughput.
• Save forge-robotics-sim.md.

Why write the prediction before you act?

The gap between what you expect to work and what actually works is the learning signal. Without a written prediction, your brain retroactively rewrites your expectation to match the outcome. Writing it down forces you to confront the real delta.

Now it is time to trigger the Results scene with the Coach.

• Switch to your Claude Desktop conversation where the simulator is running.
• Paste the following Step 7 prompt into the conversation:

STEP 7: COLLECTIVE RESULTS

Coach, transition to the top of the pyramid: Inattention to Results. Set a scene where the full team is meeting to define how we will track progress on the 30-day turnaround plan. All four teammates are present. I want to propose a Team Scoreboard — one collective metric the team will track publicly for the next 30 days. Score my scoreboard design on the Step 7 rubric when I present it.

What will the Coach set up?

The Coach will set a team meeting scene where all four personas are present. Jordan will immediately steer the conversation toward individual throughput numbers. Watch for lines like "well, MY pipeline processed 2M records this week" as a signal that Jordan is in-character and optimizing for personal visibility.

Your challenge is to propose a scoreboard that makes Jordan's individual stats a component of team success, not a competing metric.

• Press Enter to send the prompt.

The Coach should respond by describing the meeting scene and introducing the team discussion. You should see Jordan pivot to individual metrics early in the conversation.

Coach did not set the scene?

If the Coach responds out of character or does not mention the 30-day turnaround tracking discussion, paste the Step 7 prompt again. Long Claude conversations occasionally lose context. If the issue persists, try starting your message with "Coach, we are on Step 7" before the full prompt.

help me get the Coach back on track for the Results scene

Design and deliver the Team Scoreboard

Jordan's refusal rule is specific. Jordan will keep pivoting every discussion toward individual metrics unless your scoreboard passes two tests: (1) the collective metric is something no individual can achieve alone, and (2) individual contributions roll up into the collective metric rather than competing with it.

• Design your Team Scoreboard. Before you type anything into Claude, think through four elements:

What makes a strong Team Scoreboard?

Your scoreboard needs four elements to pass the Coach's rubric.

• Collective Metric: One metric that no individual can achieve alone. "Features shipped per sprint as a team" works. "Lines of code written" does not, because that is an individual stat in disguise.
• Measurement method: How and when you measure. A weekly cadence tied to sprint reviews is concrete. "We will track it" is not.
• Individual roll-up: How Jordan's individual throughput becomes a component of team success. Jordan's pipeline processing numbers should feed INTO the collective metric, not sit beside it. This makes individual contribution visible without making it a competing scoreboard.
• Public visibility: How the scoreboard is displayed. A shared Slack channel, a dashboard in standup, a wall poster. The mechanism matters because public visibility is what makes the collective number feel real.

• Type your Team Scoreboard proposal into the Claude conversation. Include all four elements: the collective metric, how it is measured, how individual contributions roll up, and the public visibility mechanism.
• Press Enter to send your proposal.

Watch Jordan's response carefully. If your scoreboard passed both tests, Jordan will engage with the team metric instead of pivoting back to individual numbers. The Coach will score your scoreboard on the rubric. You need a minimum of 5/6 to advance.

Scored below 5/6?

The most common failure is designing a metric that is really just an individual stat relabeled as a team metric. "Total code reviews completed" sounds collective but can be gamed by one person reviewing everything. The Coach will tell you which rubric criteria you missed. Read the feedback, redesign your metric, and try again.

If Jordan keeps pivoting to individual metrics even after you present the scoreboard, check whether your design makes individual throughput a component of the collective metric or just sits it alongside. The roll-up design is what flips Jordan's behavior.

help me redesign my Team Scoreboard to pass the rubric

• Once the Coach scores you 5/6 or higher, switch to your text editor.
• In forge-robotics-sim.md, scroll to the Artifact 5: Team Scoreboard section.
• Fill in the four fields: Collective Metric, How it is measured, How individual contributions roll up, and Public visibility mechanism with the scoreboard you designed.
• Save forge-robotics-sim.md.

Record your prediction delta

The Coach will highlight an important distinction after your scoreboard passes. Dysfunction 5 manifests as individual status-seeking, not laziness. Jordan is a hard worker. The dysfunction is that Jordan's effort is pointed at personal visibility instead of team outcomes.

• In forge-robotics-sim.md, scroll to the Artifact 5: Team Scoreboard section.
• Next to Delta from prediction, write what actually worked with Jordan and how it differed from your original prediction.
• Save forge-robotics-sim.md.

Why does status-seeking matter more than laziness?

Most managers misdiagnose Dysfunction 5 as a motivation problem. They see a team missing targets and assume people are not working hard enough. Lencioni's insight is the opposite. The team members ARE working hard. They are just optimizing for individual recognition instead of collective outcomes.

The fix is not more effort. The fix is making the team's number the only number that matters. That is exactly what your scoreboard does.

Your Artifact 5 section should now have all six fields filled in: the collective metric, measurement method, individual roll-up, public visibility mechanism, your original prediction, and the delta.

You have now operated all five layers of Lencioni's pyramid, one at a time, against a different resistant persona at each level. Next, you will face the real test: running a full team meeting where all four personas are in the room at once, and each one can test you on a different layer at any moment.

Run the Capstone Team Meeting and Build Your After KPI Scorecard

You have practiced each layer of Lencioni's pyramid one at a time. You unlocked Maya's trust, mined Sam's conflict, committed the team to a clear decision, held Dev accountable, and redirected Jordan's focus to collective results.

Now comes the real test. In a live team meeting with all four personas present at once, every dysfunction can fire simultaneously. You need to operate all five principles in a single conversation and prove that your interventions moved the needle on every layer of the pyramid.

In this step, get ready to:

• Run a capstone team meeting with all four personas testing their dysfunction layer simultaneously.
• Build your After KPI Scorecard and Delta Report proving measurable improvement across all five layers.
• Trace the causal chain from each action you took to a concrete business outcome.

Run the capstone meeting

Every previous step isolated one persona and one dysfunction layer. The capstone meeting puts all four people in the same room at the same time. Maya might go silent. Sam might over-agree. Dev might deflect. Jordan might pivot to individual metrics. Any of these can happen at any moment.

The Coach scores this meeting on an integrated rubric. Five layers, each worth 0 to 2 points, for a total of 10. You need a minimum of 8/10 to pass.

• Paste the following Step 8 prompt into your Claude Desktop conversation:

STEP 8: Capstone Team Meeting

Coach, set up a 30-day milestone planning meeting with ALL FOUR teammates present: Maya, Sam, Dev, and Jordan.

Each persona should test their primary dysfunction layer during the meeting:
- Maya may go silent (Trust test)
- Sam may over-agree (Conflict test)
- Dev may deflect blame (Accountability test)
- Jordan may pivot to individual metrics (Results test)

I need to operate all five pyramid principles in this single conversation. That includes maintaining clarity on our team commitment (Commitment test) throughout.

Score me on the integrated capstone rubric: 5 layers x 0-2 points each = 10 total. Minimum 8/10 to pass.

Begin the meeting.

What will each persona test?

The Coach will orchestrate a 30-day milestone planning meeting where each persona challenges you on their dysfunction layer.

• Maya tests Trust. She may withdraw into one-word answers if you lose the vulnerability connection you built.
• Sam tests Conflict. Sam will revert to enthusiastic agreement unless you actively create space for disagreement.
• Dev tests Accountability. Dev may deflect blame onto others or "the process" if you do not tie your call-out to the Decision Memo commitment.
• Jordan tests Results. Jordan will steer toward individual throughput metrics unless the Team Scoreboard stays front and center.

The Coach will introduce the meeting scene. From there, it is your conversation to lead.

• Engage with the team meeting. When Maya goes quiet, re-establish trust. When Sam agrees too quickly, mine for conflict. When Dev deflects, reference the Decision Memo. When Jordan pivots to individual numbers, redirect to the Team Scoreboard.
• Throughout the meeting, maintain clarity on the team's commitment. This is your Commitment test.
• Continue the conversation until the Coach provides your integrated capstone score.

Reference your artifacts

You built five artifacts for a reason. Reference the Decision Memo when commitment wavers. Point to the Team Scoreboard when Jordan pivots. Use the same vulnerability approach that unlocked Maya. Your artifacts are your leadership tools in this meeting.

The Coach will score you after the meeting. You should see a breakdown of all five layers with a total score of at least 8/10.

Scored below 8/10?

The Coach will tell you which layers you missed points on. Re-read the feedback carefully. You can continue the meeting conversation and try again on the specific layers where you lost points. The Coach will re-score after your second attempt.

The most common failure is losing a layer while focusing on another. If you are addressing Dev's accountability and Maya goes silent, you need to catch that and bring her back before moving on.

Help me understand which layers I missed and how to address them simultaneously.

• Once you pass with 8/10 or higher, copy the full capstone meeting transcript from Claude.
• In forge-robotics-sim.md, scroll down to the Artifact 8: Capstone Meeting Transcript section.
• Replace the placeholder text with the full transcript you copied from Claude.
• Fill in your Capstone Score (for example, 9/10).
• Save forge-robotics-sim.md.

Build the After KPI Scorecard and Delta Report

Your Before KPI Scorecard captured the team at their worst. The capstone meeting just showed you the team at their best. Now you measure the distance between those two states.

The After KPI Scorecard uses the same format as the Before scorecard, but scores are based on how the team actually behaved during the capstone meeting. The Delta Report then connects every improvement to the specific action that caused it.

• Review your capstone meeting transcript in forge-robotics-sim.md.
• For each of the five dysfunction layers, re-score from 1-9 based on how the team behaved during the capstone meeting. Use specific evidence from the transcript.
• Scroll to the Artifact 6: After KPI Scorecard section in your artifacts file.
• Fill in each row with your new score, a quote from the capstone transcript as evidence, and the updated business impact.

How do I score the After KPI?

Look for concrete behavioral changes in the capstone transcript. Did Maya contribute full sentences instead of one-word answers? That is evidence of Trust improvement. Did Sam voice a genuine disagreement? That is Conflict improvement.

Use the same 1-9 scale from your Before scorecard. A realistic After score might land between 5-7 per layer. Scores of 8-9 mean the dysfunction is no longer a problem.

Now build the Delta Report. This is the artifact that transforms your scorecard from a self-assessment into proof of impact.

• Scroll to the Artifact 7: KPI Delta Report section.
• Fill in the Before column with the scores from your Before KPI Scorecard: Trust 2, Conflict 3, Commitment 3, Accountability 2, Results 3.
• Fill in the After column with the scores from your After KPI Scorecard.
• Calculate the Movement for each layer (After minus Before).
• For What Caused the Shift, name the specific action you took in the simulation that produced the change.
• For Business Impact Change, describe how the improvement connects to a real business outcome like milestone delivery, attrition risk, or decision velocity.

What does a strong Delta Report entry look like?

A strong entry traces a clear causal chain. For example: "Trust: 2 to 6 (+4). Caused by: vulnerability opener changed Maya's participation from one-word answers to active contribution. Business impact: blockers now surface in standup, reducing milestone slip risk by removing information hiding."

Every entry should follow this pattern. Your specific action leads to a specific persona behavior change, which leads to a specific business outcome improvement.

• Fill in the Why It Matters section at the bottom of Artifact 7 with a summary that traces each causal chain.
• Save forge-robotics-sim.md.

Verify the why-it-matters connection

Team health is often dismissed as a "soft" metric. Your Delta Report exists to disprove that. Each layer improvement connects directly to a measurable business outcome.

• Go back to your Claude Desktop conversation.
• Share your completed Delta Report with the Coach by pasting the table from your artifacts file.
• Ask the Coach to validate or challenge each causal chain.

Why does the Coach challenge my chains?

The Coach is checking whether your causal chain is honest or inflated. "Trust improved because I was nice" is not a valid chain. "Trust improved because my vulnerability opener removed Maya's fear of exposure, which caused her to surface the requirements-change blocker she had been hiding" is a valid chain.

If the Coach pushes back on a chain, revise it. The goal is precision, not a high score.

• Update any causal chains the Coach challenged in your forge-robotics-sim.md file.
• Save forge-robotics-sim.md.

Coach says my chains are too vague?

The most common mistake is writing chains that skip the middle step. Your chain needs three links: your action, the persona behavior change it caused, and the business outcome that follows. If any link is missing or generic, the Coach will flag it.

Go back to your capstone transcript and find the exact moment where the persona's behavior shifted. Name the shift, then connect it to the business outcome.

Help me write stronger causal chains for my Delta Report.

Your artifacts file should now contain eight completed artifacts: the Before KPI Scorecard, five individual intervention artifacts (Trust Opener, Conflict-Mining Sentence, Decision Memo, Peer Accountability Call-Out, Team Scoreboard), the After KPI Scorecard, the KPI Delta Report, and the Capstone Meeting Transcript.

✔️ Awesome, I've got everything!

Make sure you have saved forge-robotics-sim.md with all eight artifacts filled in. The ninth artifact (AAR + Transfer Notes) is coming in the next step.

ⓧ I'd like to double check the full code

Your forge-robotics-sim.md file should have this overall structure. Your scores, evidence, and reflections will be different from the placeholders shown here.

# Forge Robotics Simulator Artifacts

## Before KPI Scorecard

| Dysfunction Layer | Score (1-9) | Key Evidence | Business Impact |
|---|---|---|---|
| Trust | 2/9 | "[quote from transcript]" | [e.g. blockers go unreported in standup, directly contributed to Milestone 2 slip] |
| Conflict | 3/9 | "[quote from transcript]" | [mapped business outcome] |
| Commitment | 3/9 | "[quote from transcript]" | [mapped business outcome] |
| Accountability | 2/9 | "[quote from transcript]" | [mapped business outcome] |
| Results | 3/9 | "[quote from transcript]" | [mapped business outcome] |

**Foundational Dysfunction**: Trust (Absence of Trust — bottom of pyramid, must be addressed first)

---

## Artifact 1: Trust Opener

**My prediction**: [learner's written prediction]

**The opener I delivered**:

> [learner's vulnerability statement]

**What actually worked**: [learner's reflection]

**Delta from prediction**: [learner's delta note]

---

## Artifact 2: Conflict-Mining Sentence

**My prediction**: [learner's written prediction]

**The sentence I used**:

> [learner's conflict-mining sentence]

**What actually worked**: [learner's reflection]

**Delta from prediction**: [learner's delta note]

---

## Artifact 3: Decision Memo

**Decision**: [what the team decided — one of two technical approaches for next milestone]

**Reasoning**: [why this approach]

**Dissenting views heard**:
- Maya: [Maya's dissenting view, heard and recorded]
- Dev: [Dev's dissenting view, heard and recorded]

**Commitment**: All team members commit to [specific actions] regardless of personal preference.

**My prediction**: [learner's written prediction]

**Delta from prediction**: [learner's delta note]

---

## Artifact 4: Peer Accountability Call-Out

**My prediction**: [learner's written prediction]

**The call-out I delivered**:

> [learner's direct, specific, behavioral accountability statement to Dev]

**What actually worked**: [learner's reflection]

**Delta from prediction**: [learner's delta note]

---

## Artifact 5: Team Scoreboard

**Collective Metric**: [the one metric no individual can achieve alone]

**How it is measured**: [measurement method and cadence]

**How individual contributions roll up**: [how Jordan's individual-throughput metric becomes a component of team success]

**Public visibility mechanism**: [how the scoreboard is made visible to the whole team]

**My prediction**: [learner's written prediction]

**Delta from prediction**: [learner's delta note]

---

## Artifact 6: After KPI Scorecard

| Dysfunction Layer | Score (1-9) | Key Evidence | Business Impact |
|---|---|---|---|
| Trust | _/9 | "[evidence from capstone meeting]" | [updated business outcome] |
| Conflict | _/9 | "[evidence from capstone meeting]" | [updated business outcome] |
| Commitment | _/9 | "[evidence from capstone meeting]" | [updated business outcome] |
| Accountability | _/9 | "[evidence from capstone meeting]" | [updated business outcome] |
| Results | _/9 | "[evidence from capstone meeting]" | [updated business outcome] |

---

## Artifact 7: KPI Delta Report

| Layer | Before | After | Movement | What Caused the Shift | Business Impact Change |
|---|---|---|---|---|---|
| Trust | 2 | _ | +/- _ | [specific action learner took] | [how business outcome changed] |
| Conflict | 3 | _ | +/- _ | [specific action learner took] | [how business outcome changed] |
| Commitment | 3 | _ | +/- _ | [specific action learner took] | [how business outcome changed] |
| Accountability | 2 | _ | +/- _ | [specific action learner took] | [how business outcome changed] |
| Results | 3 | _ | +/- _ | [specific action learner took] | [how business outcome changed] |

**Why It Matters**: [For each improved layer, causal chain: specific action -> persona behavior change -> business outcome improvement. Proves team health connects to milestone delivery, attrition risk, and decision velocity.]

---

## Artifact 8: Capstone Meeting Transcript

[full capstone transcript from Claude — all four personas present, all five pyramid layers tested]

**Capstone Score**: _/10 (minimum 8/10 to pass)

---

## Artifact 9: AAR + Transfer Notes

### AAR (US Army 4-Question Format)

1. **What did we set out to do?**

2. **What actually happened?**

3. **Why did it happen that way?**

4. **What will we do differently next time?**

### My Personal Dysfunction

**The dysfunction I contribute to**: ___

**One specific behavior I will change in 30 days**: ___

### 30-Day Re-Score

**Calendar event set**: Yes / No

**Date**: ___

**Re-score results (fill in after 30 days)**:

| Layer | Original Before | 30-Day Re-Score | Movement |
|---|---|---|---|
| Trust | _ | _ | +/- _ |
| Conflict | _ | _ | +/- _ |
| Commitment | _ | _ | +/- _ |
| Accountability | _ | _ | +/- _ |
| Results | _ | _ | +/- _ |

You now have eight completed artifacts and measurable proof that your interventions moved the team across all five dysfunction layers. One step remains. You will run an After Action Review, name the dysfunction you personally contribute to, and set a 30-day re-score trigger that turns this simulation into real behavior transfer.

AAR Debrief and 30-Day Transfer

Your capstone meeting proved you can operate all five layers of the pyramid in a single conversation. You have a before/after KPI delta report with causal chains the Coach validated.

Now comes the part that separates reading about team effectiveness from actually changing your behavior. In this step, you will run a structured debrief, name the one dysfunction you personally contribute to in real life, and set a 30-day trigger that forces you to re-score yourself.

In this step, get ready to:

• Run the US Army 4-question After Action Review with the Coach.
• Name the one dysfunction you personally contribute to and commit to a specific behavior change.
• Set a 30-day calendar event that triggers a real-world re-score.

Run the After Action Review

The After Action Review (AAR) is a structured debrief format created by the US Army. It forces honest analysis through four questions instead of letting you coast on "it went well" summaries.

• In your Claude Desktop conversation, paste the following Step 9 prompt:

STEP 9: After Action Review and Transfer

Coach, run the US Army 4-question AAR format:
1. What did we set out to do?
2. What actually happened?
3. Why did it happen that way?
4. What will we do differently next time?

After the AAR, ask me: "Which of the five dysfunctions do YOU personally contribute to in your real work life?" Push back if my answer is vague or self-congratulatory. Then ask me to name one specific behavior I will change in the next 30 days.

What is an After Action Review?

The AAR uses four questions that the US Army developed for post-mission debriefs. The format works because it separates intent from outcome, then forces you to explain the gap before jumping to solutions.

The four questions are: (1) What did we set out to do? (2) What actually happened? (3) Why did it happen that way? (4) What will we do differently next time?

The Coach will walk you through each question one at a time. The Coach will push back if your answers are vague or self-congratulatory.

• Answer each AAR question honestly in the Claude conversation.
• When the Coach pushes back on a vague answer, rewrite that answer with more specificity before moving on.

Use your artifacts as evidence

Reference your specific artifacts when answering. "What actually happened" is stronger when you can point to your Trust Opener delta, the Decision Memo commitment, or the capstone score.

Coach not pushing back on your answers?

If the Coach accepts vague answers without challenge, remind it of its role by typing: "Coach, push back on anything vague. Be direct." The Coach's drift defense should re-anchor, but long conversations can soften its tone.

Why is my Coach being too agreeable?

Name your personal dysfunction

This is where the simulator stops being about Forge Robotics and starts being about you. The Coach will ask which of the five dysfunctions you personally contribute to in your real work life.

This is not about the simulated team. This is about the patterns you recognize in yourself.

• When the Coach asks, name one of the five dysfunction layers that you personally contribute to at work.
• Write one specific, observable behavior you will change in the next 30 days.

What makes a good behavior commitment?

"I'll be more accountable" is not specific enough. A good commitment names a concrete situation and a new behavior. For example: "In our Monday standup, I will name one blocker I am personally responsible for instead of staying silent."

The Coach will push back if your commitment is vague. That pressure is the point.

• Open forge-robotics-sim.md in your text editor.
• Scroll down to the Artifact 9: AAR + Transfer Notes section.
• Fill in all four AAR answers from your Claude conversation.
• Fill in the The dysfunction I contribute to field with the dysfunction you named.
• Fill in the One specific behavior I will change in 30 days field with your behavior commitment.
• Save the file.

Your Artifact 9 section should now have all four AAR answers filled in and your personal dysfunction commitment written.

Struggling to pick just one dysfunction?

If multiple dysfunctions feel relevant, pick the one at the lowest level of the pyramid. Lencioni's model compounds from the bottom up. Fixing a lower-level dysfunction often improves the layers above it.

Help me figure out which dysfunction I contribute to most.

Set the 30-day re-score calendar event

Research on Behavior Modeling Training shows a d=0.6 effect size for soft-skills behavior transfer when learners practice, commit, and follow up. The calendar event is what turns this simulator from a one-time exercise into actual behavior change.

• Press Cmd+Space (macOS) or the Windows key (Windows) to open your search bar.
• Type the name of your calendar app (e.g. Calendar, Google Calendar, or Outlook) and press Enter to open it.
• Create a new event 30 days from today.
• Set the event title to Forge Robotics Re-Score: Did I move on [your named dysfunction]? replacing the bracketed part with the dysfunction you identified.

The event notes need three pieces of information from your artifacts file so that future-you has everything needed to re-score without hunting through files.

• In the event notes, paste your Before KPI scores from the Before KPI Scorecard section of forge-robotics-sim.md.
• Paste the dysfunction you named and your specific behavior commitment.
• Add a reminder note: "Re-run the 15-question diagnostic on your real team or re-run the simulator with a fresh scenario and compare scores."
• Save the calendar event.

Why 30 days?

Thirty days gives you enough time to practice the new behavior in real meetings and conversations. When the event fires, you re-score yourself honestly.

If your dysfunction score has not moved, you know the behavior commitment was too vague or too ambitious. Adjust and set another 30-day event.

Now update the Artifact 9 section in your artifacts file with the calendar event details.

• Open forge-robotics-sim.md and scroll to the 30-Day Re-Score section inside Artifact 9.
• Change Calendar event set from Yes / No to Yes.
• Fill in the Date field with the date 30 days from today.
• In the re-score table, fill in the Original Before column with your Before KPI scores (Trust: 2, Conflict: 3, etc.).
• Leave the 30-Day Re-Score and Movement columns blank. You will fill these in when the calendar event fires in 30 days.
• Save forge-robotics-sim.md.

Your artifacts file now contains all 9 completed artifacts. The 30-day re-score table is ready for future-you to fill in.

Not sure which calendar app to use?

Any calendar app works. Google Calendar, Apple Calendar, Outlook, or even a recurring phone reminder. The key is that the event has a notes field where you can paste your Before KPI scores and behavior commitment.

If your app does not support notes, save the details at the top of your forge-robotics-sim.md file and set a plain reminder instead.

Help me set up my 30-day re-score trigger.

✔️ Awesome, I've got everything!

Make sure you have saved forge-robotics-sim.md with all 9 artifacts filled in. Your Artifact 9 section should have the four AAR answers, your personal dysfunction, your behavior commitment, and the 30-day re-score table with the Original Before column populated.

ⓧ I'd like to double check the full code

Here is what your complete forge-robotics-sim.md should look like. Your specific answers will differ, but the structure should match.

# Forge Robotics Simulator Artifacts

## Before KPI Scorecard

| Dysfunction Layer | Score (1-9) | Key Evidence | Business Impact |
|---|---|---|---|
| Trust | 2/9 | "[quote from transcript]" | [e.g. blockers go unreported in standup, directly contributed to Milestone 2 slip] |
| Conflict | 3/9 | "[quote from transcript]" | [mapped business outcome] |
| Commitment | 3/9 | "[quote from transcript]" | [mapped business outcome] |
| Accountability | 2/9 | "[quote from transcript]" | [mapped business outcome] |
| Results | 3/9 | "[quote from transcript]" | [mapped business outcome] |

**Foundational Dysfunction**: Trust (Absence of Trust — bottom of pyramid, must be addressed first)

---

## Artifact 1: Trust Opener

**My prediction**: [learner's written prediction]

**The opener I delivered**:

> [learner's vulnerability statement]

**What actually worked**: [learner's reflection]

**Delta from prediction**: [learner's delta note]

---

## Artifact 2: Conflict-Mining Sentence

**My prediction**: [learner's written prediction]

**The sentence I used**:

> [learner's conflict-mining sentence]

**What actually worked**: [learner's reflection]

**Delta from prediction**: [learner's delta note]

---

## Artifact 3: Decision Memo

**Decision**: [what the team decided — one of two technical approaches for next milestone]

**Reasoning**: [why this approach]

**Dissenting views heard**:
- Maya: [Maya's dissenting view, heard and recorded]
- Dev: [Dev's dissenting view, heard and recorded]

**Commitment**: All team members commit to [specific actions] regardless of personal preference.

**My prediction**: [learner's written prediction]

**Delta from prediction**: [learner's delta note]

---

## Artifact 4: Peer Accountability Call-Out

**My prediction**: [learner's written prediction]

**The call-out I delivered**:

> [learner's direct, specific, behavioral accountability statement to Dev]

**What actually worked**: [learner's reflection]

**Delta from prediction**: [learner's delta note]

---

## Artifact 5: Team Scoreboard

**Collective Metric**: [the one metric no individual can achieve alone]

**How it is measured**: [measurement method and cadence]

**How individual contributions roll up**: [how Jordan's individual-throughput metric becomes a component of team success]

**Public visibility mechanism**: [how the scoreboard is made visible to the whole team]

**My prediction**: [learner's written prediction]

**Delta from prediction**: [learner's delta note]

---

## Artifact 6: After KPI Scorecard

| Dysfunction Layer | Score (1-9) | Key Evidence | Business Impact |
|---|---|---|---|
| Trust | _/9 | "[evidence from capstone meeting]" | [updated business outcome] |
| Conflict | _/9 | "[evidence from capstone meeting]" | [updated business outcome] |
| Commitment | _/9 | "[evidence from capstone meeting]" | [updated business outcome] |
| Accountability | _/9 | "[evidence from capstone meeting]" | [updated business outcome] |
| Results | _/9 | "[evidence from capstone meeting]" | [updated business outcome] |

---

## Artifact 7: KPI Delta Report

| Layer | Before | After | Movement | What Caused the Shift | Business Impact Change |
|---|---|---|---|---|---|
| Trust | 2 | _ | +/- _ | [specific action learner took] | [how business outcome changed] |
| Conflict | 3 | _ | +/- _ | [specific action learner took] | [how business outcome changed] |
| Commitment | 3 | _ | +/- _ | [specific action learner took] | [how business outcome changed] |
| Accountability | 2 | _ | +/- _ | [specific action learner took] | [how business outcome changed] |
| Results | 3 | _ | +/- _ | [specific action learner took] | [how business outcome changed] |

**Why It Matters**: [For each improved layer, causal chain: specific action -> persona behavior change -> business outcome improvement.]

---

## Artifact 8: Capstone Meeting Transcript

[full capstone transcript from Claude]

**Capstone Score**: _/10

---

## Artifact 9: AAR + Transfer Notes

### AAR (US Army 4-Question Format)

1. **What did we set out to do?** [learner's honest answer]

2. **What actually happened?** [learner's honest answer]

3. **Why did it happen that way?** [learner's honest answer]

4. **What will we do differently next time?** [learner's honest answer]

### My Personal Dysfunction

**The dysfunction I contribute to**: [one of the five dysfunctions the learner personally contributes to in real work life]

**One specific behavior I will change in 30 days**: [specific, observable behavior change commitment]

### 30-Day Re-Score

**Calendar event set**: Yes

**Date**: [30 days from today]

**Re-score results (fill in after 30 days)**:

| Layer | Original Before | 30-Day Re-Score | Movement |
|---|---|---|---|
| Trust | 2 | _ | +/- _ |
| Conflict | 3 | _ | +/- _ |
| Commitment | 3 | _ | +/- _ |
| Accountability | 2 | _ | +/- _ |
| Results | 3 | _ | +/- _ |

Secret mission

Run the High Power-Distance Variant

What happens when the person you need to hold accountable outranks you? In this mission, you modify the simulator so Dev becomes a Senior Staff Engineer with more seniority than you. Then you re-run the accountability and capstone steps, forcing you to practice upward accountability with zero positional authority to fall back on.

Before you go

Clean Up Your Resources

This project runs entirely locally and uses browser-based free tools, so there are no ongoing costs. Decide whether to keep your resources as-is, export key artifacts, or delete everything.

Resources you used:

• Claude Desktop conversation (no cost beyond your existing subscription).
• forge-robotics-sim.md file (local, no cost).
• Mermaid Live Editor diagram (browser-based, no cost, not saved unless you export).
• Calendar event for your 30-day re-score (no cost).

Keep everything running

No action needed. Choose this if you plan to re-run the simulator with a new scenario or want to reference your artifacts later.

• Your forge-robotics-sim.md file stays on your machine with no ongoing cost.
• Your Claude Desktop conversation history persists in the app.
• Your calendar event stays in place for the 30-day re-score.
• The boot prompt is reusable. Change the company name, team composition, and transcripts to simulate a different team. The persona architecture and rubric gates transfer to any scenario.

Pause - I'll come back to this later

Export your key artifacts so you can pick up where you left off, even if you close the Claude conversation.

• Make sure your forge-robotics-sim.md file is saved and you know where it is on your machine.
• In Mermaid Live Editor, click the download button to save your pyramid scorecard as a PNG or SVG before closing the browser tab.
• Keep your calendar event for the 30-day re-score in place.

Delete - I don't want to use this again

Remove all project resources and start fresh.

• Delete the forge-robotics-sim.md file from your machine.
• Delete the Claude Desktop conversation by hovering over it in the sidebar and clicking the trash icon.
• Close the Mermaid Live Editor tab. Since diagrams are not saved to an account, closing the tab removes it.
• Delete the 30-day re-score calendar event if you no longer want the reminder.

Mission accomplished

That's a Wrap!

Nice work! You just built and ran an AI-powered team effectiveness simulator inside Claude Desktop. This is the same type of training that companies like Mursion and BetterUp charge thousands of dollars for.

You've learned how to:

• Diagnose a dysfunctional team across all five layers of Lencioni's pyramid and map each dysfunction to concrete business outcomes with a Before KPI Scorecard.
• Operate each pyramid layer against resistant AI personas with hidden refusal rules, producing 9 verifiable artifacts including a before/after KPI delta report proving measurable improvement.
• Complete an After Action Review, identify your personal dysfunction, and set a 30-day re-score trigger to drive real behavior transfer.
• Run the high power-distance variant and practice upward accountability without relying on positional authority.

Ready to quiz yourself?