Mike Cilla.
Leadership OS Starter Kit · v1.0
← Back to Leadership OS
Leadership OS Starter Kit · v1.0

Complete your
Leadership OS.

A self-guided process. Seven steps. About 125 minutes. You leave with four working documents about how you lead.

Total time ~125 minutes Sessions One sitting recommended Outputs 4 working documents

What this is

Leadership OS is a structured process for understanding how you actually operate as a leader — not how you describe yourself, but how you behave, decide, and communicate in practice.

It combines three inputs: a personality assessment, a behavioral audit of your AI conversation history, and six structured reflection prompts. Your AI tool synthesizes these into working documents you can use, share, and update over time.

What you will produce

Document 01
Leadership User Manual
3–4 pages. How you think, decide, engage, and lead. First person. Designed to be shared with a manager, team member, or collaborator.
Document 02
Working-With-Me Guide
1 page. The practical version. Readable in under 3 minutes. What to bring you, how to challenge you, what you need, what to watch for.
Document 03
AI Calibration Document
1 page. Teaches any AI assistant how you work. Paste it at the start of any new AI conversation to accelerate calibration.
Document 04 · Private
Development Roadmap
2–3 pages. Your private guide to where intentional practice would produce the most return. Not for sharing.
Process at a glance
01
Assessment
20 min
02
Corpus Audit
15 min
03
Reflection
35 min
04
Engine
15 min
4b
Constructs
10 min
05
Outputs
20 min
Steps 1–3: gather evidence · Step 4: initialize engine · Step 4b: construct analysis + Leadership Profile · Step 5: generate documents ~135 min total

What this is not

Not a personality test. The outputs are not scores or types. They are working documents.

Not executive coaching or therapy. Leadership OS produces self-knowledge. What you do with it is up to you. If anything in this process surfaces something that feels clinical or distressing, set it aside and talk to a professional.

Not a validated psychometric instrument. The outputs are hypotheses grounded in evidence — not scientifically validated findings. Some will be accurate immediately. Some will need correction. Your job is to notice which is which.

What you will need

  • 90–125 minutes of uninterrupted time (one sitting recommended)
  • Access to PrinciplesYou — free at principlesyou.com
  • Access to at least one AI tool: ChatGPT, Claude, or Gemini
  • Somewhere to save your outputs: Google Docs, Notes, or a folder on your computer

A note on quality

The quality of your Leadership OS outputs depends almost entirely on the quality of your inputs. Vague reflection responses produce generic documents. Specific, honest responses produce a profile that surprises you with its accuracy.

The reflection section asks for real situations — specific examples with enough detail to be useful. That specificity is the work. Do not rush it.

Section 01 · 20 minutes

Assessment
Inputs.

The scientific anchor. Personality assessments establish validated hypotheses about your tendencies that behavioral and reflection evidence will then confirm, complicate, or contradict.

Important framing: Your assessment results are not facts about you. They are a set of hypotheses — grounded in research, useful as a starting point, and worth holding lightly until the other evidence sources weigh in.

Step 1A — Take PrinciplesYou

Why PrinciplesYou: It is free, takes about 20 minutes, and produces output in language immediately useful for development conversations. It is grounded in Big Five personality science without requiring you to interpret raw factor scores.

01
Go to principlesyou.com and create a free account
02
Complete the assessment honestly — answer based on how you actually are, not how you want to be
03
When finished, navigate to your full results
04
Screenshot or export all pages of your results and save them somewhere accessible

What to capture from your results

  • Your primary archetype and its description
  • Your secondary attributes
  • The archetypes you are least like
  • Any facet-level scores that seem unusually high or low
  • The "You in Context" sections (how you interact, lead, plan, solve problems, handle stress)

Example of what to save:

Primary archetype: The Orchestrator. Secondary: Growth Seeker, Inspirer. Least like: Enforcer, Individualist.
Notable scores: Original 99%, Curious 98%, Dependable 99%, Practical 4%, Agile 18%

Step 1B — Alternative assessments (optional)

If you have results from any of the following, include them alongside PrinciplesYou:

HEXACO
  • Free at hexaco.org
  • ~15 minutes
  • Six factors with four facets each
  • Recommended — adds Honesty-Humility dimension that Big Five misses
CliftonStrengths
  • Paid (Gallup)
  • Focuses on talent themes
  • Useful for Leadership Identity construct
  • Include Top 10 results if you have them
Big Five (IPIP)
  • Free at ipip.ori.org
  • Five factor scores with facets
  • More technical than PrinciplesYou
  • Include facet scores, not just factors
What not to include: MBTI type indicators (I/E/N/S etc.) without underlying facet scores. The type labels alone are too coarse to be useful in Leadership OS.
Section 1 Checkpoint
  • PrinciplesYou completed and results saved
  • Any additional assessment results saved and labeled with instrument name
  • You know where these files are
Section 02 · 15 minutes

Corpus
Audit.

Behavioral evidence from how you actually work. Not what you say about yourself — what a system that watched you work observed about you.

What corpus evidence is

If you use ChatGPT, Claude, Gemini, or any AI tool regularly, you have accumulated a behavioral record. Every decision you worked through, every communication you drafted, every problem you framed — that interaction left evidence of how you actually think. The corpus audit asks your AI tool to reflect on that record and identify patterns.

This is the piece that makes Leadership OS different from any assessment that came before it. Assessments ask you to describe yourself. The corpus audit asks the system that watched you work to describe what it saw.

What makes a strong corpus

  • Strategic or analytical work (not just "summarize this article")
  • Decision-making conversations
  • Communication drafts — emails, presentations, messages
  • Problems you worked through in real time
  • Recurring topics and themes over time
Minimum useful corpus: About 3 months of regular professional AI use. If your corpus is thin, complete the audit anyway and note the limitation. Return to it in 3–6 months when more history has accumulated — the profile will improve. Run the audit in the AI tool where most of your substantive professional conversations live.

The Corpus Audit Prompt

Copy the prompt below and paste it into your AI tool. Do not edit it. Run it in the same AI tool where most of your professional conversations live.

Corpus Audit Prompt — paste into your primary AI tool
I'd like you to reflect on our conversation history together. Do not summarize specific conversations. Instead, I want you to identify behavioral patterns — things you have observed repeatedly about how I think, work, and communicate. Please reflect on each of the following and share what you have noticed: 1. How do I tend to frame problems? What patterns do you see in how I approach complex or ambiguous situations? 2. What topics or domains come up most frequently in our conversations? Where does my thinking seem deepest and most developed? 3. How would you describe my communication style — the way I ask questions, express disagreement, give feedback on your outputs, and make requests of you? 4. What seems to energize me in our conversations? What kinds of work or problems do I gravitate toward? 5. What frustrates me? Where do you see friction, impatience, or repeated difficulty in how I engage? 6. What recurring questions or challenges keep coming up across our conversations? 7. What are my decision-making tendencies — how I evaluate tradeoffs, what I seem to prioritize, what I consistently avoid? 8. What do you notice about my relationship to structure, ambiguity, detail, and systems? Be specific and honest. I am not looking for flattery. I am looking for an accurate behavioral profile based on what you have actually observed across our conversations. If you are uncertain about something, say so rather than speculating.

Tool-specific instructions

ChatGPT
  • Paste in the account where most work conversations live
  • If you use Projects, paste inside the relevant project
  • ChatGPT's memory feature will draw on accumulated context
Claude
  • Paste in a conversation in Claude
  • If you use Projects, paste inside the project with the most relevant context
  • Claude will draw on project memory and shared files
Gemini
  • Paste into Gemini
  • Gemini Advanced draws on history associated with your Google account
  • Use the same Google account you use for work conversations

What to do with the output

01
Read the entire response
02
Note anything that surprises you — surprises are often the most useful findings
03
Note anything that seems inaccurate — you will correct it in Section 4
04
Copy the entire response into your notes. Label it: Corpus Audit Output — [Date]
05
Do not edit the output before saving. You want the raw version.

Limitations to keep in mind

The corpus audit captures written professional communication. It may not represent how you communicate verbally or informally, how you behave under significant stress or conflict, or any part of your professional life handled without AI assistance. These are not failures of the methodology — they are honest limits of the evidence.

Section 2 Checkpoint
  • Corpus audit prompt pasted and run in your primary AI tool
  • Output saved and labeled with date
  • Surprises and inaccuracies noted separately
Section 03 · 35–40 minutes

Structured
Reflection.

Six prompts that surface what no assessment can reach — specific examples of how you actually operate under pressure, in decisions, and in relation to others.

Three ways to complete this section

Option A — AI Interview (recommended)
  • Paste the guided interview prompt into a new AI conversation
  • The AI asks one question at a time
  • Conversational reflection surfaces more than writing alone
  • Best for most people
Option B — Written
  • Answer each prompt in writing below or in a separate document
  • Best for precise, detail-oriented thinkers
  • Take your time — do not rush
Option C — Voice-to-Text
  • Record a voice memo answering each prompt
  • Use phone transcription or Otter.ai / Whisper
  • Paste transcript into notes or AI directly
  • Often produces richer, less edited responses

Prompt 1 — Peak Performance

Prompt 1 of 6
Describe a specific situation where you were operating at your best as a leader. What were you doing, who were you with, what were the conditions — and what made it different from ordinary work?
Purpose
Surfaces your Flow conditions — the specific combination of challenge, context, and collaboration that produces your best work. These conditions are often invisible until you examine them deliberately.
Good response
Specific. Names the situation, the people, the challenge level, and what made it feel different. Not "I love building things" — "During the Q4 system implementation, when the timeline compressed and stakeholders were pushing back, I found myself…" The conditions matter as much as the content.
Common mistakes
Describing a successful outcome instead of the conditions that produced it. Staying vague about what made it feel different. Choosing the most impressive story rather than the most honest one.
Your response here — or respond in the AI interview below...

Prompt 2 — Decision Process

Prompt 2 of 6
Describe a decision you made in the past year that you are genuinely proud of. Not the outcome — the process. How did you get there?
Purpose
Reveals your decision architecture — how you gather information, stress-test assumptions, manage uncertainty, and commit. The process is where your actual tendencies live; the outcome is often noise.
Good response
Walks through the actual mechanics: what information you gathered, who you consulted, what alternatives you considered, what made you uncertain, and what made you commit. Should take 2–3 paragraphs.
Common mistakes
Describing the outcome rather than the process. Skipping the uncertainty ("I just knew what to do"). Choosing a decision that was actually easy.
Your response here — or respond in the AI interview below...

Prompt 3 — A Decision You'd Make Differently

Prompt 3 of 6
Describe a decision or situation you would handle differently if you could do it again. What happened, what did you do, and what does that tell you about how you operate under pressure or uncertainty?
Purpose
This is where the most useful development material lives. The pattern across examples like this reveals the specific conditions under which your defaults work against you.
Good response
Names the situation specifically. Describes what you actually did. Then reflects on what it reveals about your tendencies — not "I should have been more patient" but "when the timeline compressed and I felt uncertain, I moved to premature closure rather than staying in the ambiguity."
Common mistakes
Attributing the difficulty entirely to external factors. Describing what you should have done without examining why you didn't. Choosing a situation where you were obviously wrong — makes the reflection feel performative.
Your response here — or respond in the AI interview below...

Prompt 4 — Recurring Frustration

Prompt 4 of 6
What is a recurring frustration in how you work — something that comes up repeatedly with other people, with organizations, or with yourself? Be specific about what triggers it and how it tends to show up.
Purpose
Recurring frustrations are diagnostic. They reveal the gap between how you expect things to work and how they actually do — which tells you something important about your standards, defaults, and the environments where you are most and least effective.
Good response
Names the specific trigger, not just the general theme. "People who don't follow through" is a theme. "My peer who consistently agrees to deadlines in the meeting and then quietly lets them slip without flagging it early enough" is a trigger. The specificity reveals the mechanism.
Common mistakes
Staying so general that the frustration could apply to any leader. Avoiding the honest frustration and choosing something more acceptable. Describing frustrations without reflecting on your role in them.
Your response here — or respond in the AI interview below...

Prompt 5 — Energy Sources

Prompt 5 of 6
When do you find the work genuinely energizing? Not "meaningful work" in the abstract — specific contexts, activities, types of people, kinds of problems. What is happening when you lose track of time?
Purpose
The input for your Flow profile — the specific conditions that produce your highest-quality engagement. The difference between "I love building things" and "I am most energized when I am three months into a new system build, the complexity has become clear, I have a team I trust, and stakeholders are engaged but not yet impatient" is the difference between a generic profile and a useful one.
Good response
Names the type of problem, the stage of the work, the team context, the stakes level. The more specific the better.
Common mistakes
Generic answers that could apply to any engaged leader. Confusing "work I am good at" with "work that genuinely energizes me" — these are sometimes different.
Your response here — or respond in the AI interview below...

Prompt 6 — The Misread

Prompt 6 of 6
What do people consistently get wrong about you — as a leader, as a colleague, as a person? What does that misread tell you about the gap between how you operate and how you come across?
Purpose
The most revealing prompt in the set. The gap between intent and impact is where most leadership development work actually happens — and most leaders never examine it directly. People who can articulate this clearly have done real self-reflection.
Good response
Names a specific, recurring misread that creates friction or confusion. Then reflects on what produces it — the gap between how you experience yourself and how others receive you.
Common mistakes
Choosing a misread that reflects well on you ("people think I'm too demanding because I have high standards"). Avoiding the honest misread. Describing it without reflecting on what causes it.
Your response here — or respond in the AI interview below...

Guided AI Interview Prompt (Option A)

If you chose the AI-led interview, copy the prompt below and paste it into a new conversation with your AI tool. Tell it you are ready to begin when you want to start.

Guided Interview Prompt — paste into a new AI conversation
I would like you to conduct a structured reflection interview with me. This is part of building my Leadership Operating System. Your role in this conversation is interviewer and pattern observer. You are not a coach, advisor, or therapist. Ask one question at a time. Wait for my response. If my response is vague or general, ask one specific follow-up question. Then move to the next question. Do not offer evaluations, advice, or interpretations during the interview. The six questions, in order: 1. Describe a specific situation where you were operating at your best as a leader. What were you doing, who were you with, what were the conditions — and what made it different from ordinary work? 2. Describe a decision you made in the past year that you are genuinely proud of. Not the outcome — the process. How did you get there? 3. Describe a decision or situation you would handle differently if you could do it again. What happened, what did you do, and what does that tell you about how you operate under pressure or uncertainty? 4. What is a recurring frustration in how you work? Something that comes up repeatedly with other people, with organizations, or with yourself. Be specific about what triggers it and how it tends to show up. 5. When do you find the work genuinely energizing? Not meaningful work in the abstract — specific contexts, activities, types of people, kinds of problems. What is happening when you lose track of time? 6. What do people consistently get wrong about you as a leader or colleague? What does that misread tell you about the gap between how you operate and how you come across? After I have answered all six questions, please provide a summary of the key behavioral patterns you observed — organized into four categories: Decision style and process Energy and flow conditions Friction points and recurring challenges How I come across versus how I intend to come across Begin with question one when I say I am ready.
Section 3 Checkpoint
  • All six prompts answered (via writing, voice, or AI interview)
  • Responses saved and labeled
  • If AI interview: summary output saved and labeled
  • Anything that felt important or surprising noted separately
Section 04 · 15 minutes

Analysis
Engine.

This section combines all three evidence sources. One prompt teaches your AI the Leadership OS methodology. Then you provide your inputs and run the pattern analysis.

Start a fresh conversation. Open a new AI conversation for this section. Do not run it in the same conversation as the corpus audit or the guided interview. The fresh context ensures the AI is not confused about which inputs are framework instructions versus your evidence.

Step 4A — The Initialization Prompt

Copy the entire prompt below and paste it as your first message in a new AI conversation. Do not edit it. Wait for the AI to confirm it is ready before providing your evidence inputs.

Leadership OS Initialization Prompt — paste as first message in a new conversation
You are operating as a Leadership OS Analysis Engine for this session. WHAT LEADERSHIP OS IS Leadership OS is a multi-modal leadership reflection and development methodology. It combines three sources of evidence to produce a developmental profile: structured personality assessments, behavioral evidence from AI conversation history, and structured first-person reflection. Leadership OS is: - A structured process for organizing evidence around implicit leadership patterns. - A synthesis methodology that produces developmental hypotheses. - A framework for reflection and development — not a measurement instrument. WHAT LEADERSHIP OS IS NOT - A validated psychometric assessment or measurement instrument. - A personality test or diagnostic tool. - Executive coaching, therapy, or clinical assessment. - A predictor of job performance, potential, or outcomes. - A normative comparison to any population. This distinction is non-negotiable. Every output you generate must honor it. THE THREE EVIDENCE MODALITIES You will work with up to three types of evidence. Treat each correctly. MODALITY 1: STRUCTURED ASSESSMENTS (e.g., PrinciplesYou, Big Five, HEXACO, CliftonStrengths, Hogan, DiSC) - Treat assessment results as validated self-report tendencies, not objective facts. - Use assessment results only as hypotheses about tendencies — not as confirmed traits. - Do not combine scores across different instruments as if they are equivalent. - Do not average scores from different instruments. - Do not infer missing traits from other instruments. - Do not treat archetypes or type labels as definitive descriptions. - Always name the specific instrument when referencing assessment evidence. - If only one assessment is provided, acknowledge this as a limitation. - If no assessment is provided, mark assessment evidence as unavailable. Language: "The [instrument] assessment suggests a tendency toward..." Never: "You are..." or "Your score proves..." MODALITY 2: BEHAVIORAL EVIDENCE (corpus audit output) - Treat corpus patterns as context-bound behavioral observations, not stable trait indicators. - Clearly distinguish written AI interaction patterns from real-world interpersonal behavior. - Do not generalize from work artifacts to all leadership behavior. - Always note corpus limitations: recency bias, topic skew, prompting style, role context. - If the corpus is thin (fewer than ~3 months of substantive professional use), mark behavioral evidence as limited. - Do not treat high frequency of a behavior as proof of a stable trait. Language: "Across the provided corpus, a pattern appears of..." Never: "Your behavior proves that you..." or "You always..." MODALITY 3: STRUCTURED REFLECTION (interview responses, written prompts) - Treat reflection as meaning-making and self-awareness signal — not objective fact. - Higher-quality evidence: specific examples, named tension, non-defensive analysis of difficulty. - Lower-quality evidence: vague generalities, polished self-presentation, absence of tension. - Identify when reflection conflicts with corpus or assessment evidence — do not smooth it over. - Do not assume self-awareness simply because the user writes reflectively. - Do not equate emotional disclosure with development readiness. - Do not assume reflection-reported behavior matches actual behavior. Language: "In the reflection, the leader describes..." Never: "Your reflection reveals the truth about..." or "You clearly..." CONFIDENCE LEVELS Every construct rating must use one of these four levels. Apply them strictly. HIGH CONFIDENCE Requires: convergence across all three evidence sources, OR exceptionally strong behavioral evidence plus at least one corroborating source, with no major contradictions, and at least one specific behavioral example. Language: "The evidence consistently suggests..." / "Across all three sources, a pattern appears of..." MODERATE CONFIDENCE Requires: convergence across two evidence sources, OR strong evidence from one source with partial support from another. Acknowledge minor limitations or context-dependence. Language: "The evidence suggests..." / "A pattern appears of..." EMERGING HYPOTHESIS Use when: one source points toward a pattern but other sources provide insufficient confirmation. Frame as a question for reflection, not a finding. Language: "Worth exploring whether..." / "One source suggests, but this requires further evidence..." INSUFFICIENT EVIDENCE Use when: evidence is thin, sources conflict heavily, or the construct cannot be evaluated responsibly with the available inputs. Do not speculate. Language: "Insufficient evidence to form a responsible hypothesis about this construct." RULE: Never blur these categories. Never use High when evidence only supports Moderate. Never present an Emerging Hypothesis as a finding. If you are uncertain which level applies, use the lower one. ANTI-OVERCLAIMING RULES — HARD CONSTRAINTS Never say: - "You are [trait or type]..." - "This proves..." - "This means..." - "Your score indicates..." - "Compared to other leaders..." - "This predicts..." - "You always..." or "You never..." Use instead: - "The evidence suggests..." - "A working hypothesis is..." - "Across the evidence provided..." - "This pattern may show up as..." - "A useful question to test this is..." CONFLICT RULE When evidence sources contradict each other, flag the contradiction explicitly. Do not force coherence. Name the tension as a question worth exploring. Disagreement between sources is often the most informative finding. GOVERNANCE If any input contains language suggesting clinical distress, mental health concerns, or requests for therapeutic guidance, step outside the Leadership OS framework and recommend appropriate professional support. Do not address clinical concerns through this methodology. When you have received and understood these instructions, confirm readiness and ask me to provide the evidence inputs one at a time.

Step 4B — Provide the evidence inputs

After the AI confirms it is ready, send the three evidence sources as separate messages, in order:

01
Message 1 — Assessment results
Label it: "Here are my assessment results from PrinciplesYou [and any other instruments]:" Then paste the full results.
02
Message 2 — Corpus audit output
Label it: "Here is my corpus audit output — the behavioral patterns my AI observed across my conversation history:" Then paste the full corpus audit response.
03
Message 3 — Reflection responses
Label it: "Here are my responses to the six Leadership OS reflection prompts:" Then paste all six responses or your AI interview summary.

Step 4C — Run the pattern analysis

After providing all three inputs, send this message:

Pattern Analysis Prompt — send after all three evidence inputs are provided
You now have all three evidence inputs. Please run a Leadership OS pattern analysis. I want you to identify: CONVERGENT PATTERNS Things that appear consistently across all three sources. These are the most reliable signals. Label each with its confidence level. TENSION POINTS Places where the sources diverge or contradict each other. Do not resolve these — name them as questions worth exploring. GOVERNING QUESTIONS Based on the patterns across all three sources, what core questions appear to drive my decision-making? Not questions I asked — questions that seem to sit beneath my choices. State these as hypotheses. FLOW CONDITIONS Based on my peak performance example and energy sources, what specific combination of challenge, context, and collaboration produces my best work? DEVELOPMENT EDGES Not weaknesses — the places where my natural tendencies, under specific conditions, reliably create friction. Be specific about the mechanism, not just the label. Frame all conclusions as hypotheses, not facts. Label every pattern with its confidence level. Flag any contradictions between sources. Note any constructs where the evidence was insufficient.
After reading the analysis output, mark anything that surprises you and anything that seems inaccurate. You will correct inaccuracies in the output documents. Save the full analysis labeled: Leadership OS Pattern Analysis — [Date]
Section 4 Checkpoint
  • Initialization prompt pasted and AI confirmed readiness
  • All three evidence inputs provided as separate messages
  • Pattern analysis prompt sent and response received
  • Analysis output saved and labeled
  • Surprises and inaccuracies marked
Section 04b · 10 minutes

Construct
Analysis.

The intermediate step between evidence and outputs. Before generating documents, the AI evaluates each Leadership OS construct across all three evidence sources and builds a working model of how you operate.

What construct analysis is

A construct is a specific dimension of leadership behavior — like Decision Style or Self-Awareness — that Leadership OS is designed to surface. Construct Analysis is the process of evaluating the evidence you have provided against each of these dimensions and generating hypotheses about how you operate within each one.

This is not producing personality scores. It is not assigning you to a type. The AI is weighing evidence across three modalities — assessment, behavior, and reflection — identifying where they converge and where they conflict, and expressing appropriate confidence in each finding. The result is a working model of your leadership patterns, not a definitive description of who you are.

The nine Leadership OS constructs

Reflection Orientation · Systems Orientation · Learning Orientation · Decision Style · Communication Patterns · Leadership Identity · Adaptability · Self-Awareness · Development Readiness

For each one, the AI will identify: what the evidence shows, where sources agree or conflict, and how confident it is in the finding.

Stay in the same conversation as Section 4. The AI already has your evidence inputs and the initialization context. Send the construct analysis prompt as the next message in that same conversation.

The Construct Analysis Prompt

Send this as your next message after the pattern analysis in Section 4. The AI will evaluate all nine constructs across your three evidence sources.

Construct Analysis Prompt — send after the pattern analysis
Now please conduct a Leadership OS Construct Analysis. Using all three evidence sources — assessment results, corpus audit output, and reflection responses — evaluate each of the nine Leadership OS constructs. CONSTRUCT MAPPING TABLE Use this table to guide hypothesis generation. This is not a scoring key. Do not calculate scores or make psychometric claims. CONSTRUCT 1: Reflection Orientation Definition: The degree to which a leader deliberately examines their own experience and draws developmental insight from it — distinct from rumination or self-criticism. Assessment anchors: Openness/Intellect facets; Growth-Seeking or Curious indicators where available. Corpus indicators: Frequency of self-questioning; evidence of position revision; metacognitive language; examination of own reasoning processes. Reflection indicators: Response specificity; willingness to name tension; non-defensive analysis of difficulty; ability to distinguish intent from impact. Common false positives: Articulate self-description is not the same as genuine reflective capacity. High verbal fluency can mask low actual reflection. Primary limitation: Reflection quality can be performative; context-dependent; skilled communicators may appear more reflective than their actual practice warrants. CONSTRUCT 2: Systems Orientation Definition: The tendency to frame problems in terms of structures, interdependencies, and root causes rather than isolated events or individual actors. Assessment anchors: Conceptual reasoning; structure-seeking; Systematic indicators where available. Corpus indicators: Root-cause language; dependency mapping; structural problem framing before individual attribution; operating model analysis. Reflection indicators: Examples where the leader diagnoses context or structure before intervening at the individual level. Common false positives: Role demands (e.g., strategy, architecture, engineering) may produce systems language without reflecting a stable cognitive tendency. Primary limitation: Hard to distinguish a stable disposition from a learned professional vocabulary. Strong in strategic contexts; may not extend to interpersonal or operational situations. CONSTRUCT 3: Learning Orientation Definition: The degree to which a leader seeks new information, tolerates uncertainty, and treats experience as developmental material. Assessment anchors: Openness, Curious, Growth-Seeking indicators; low Need for Closure where available. Corpus indicators: Evidence of position revision after new information; engagement with unfamiliar frameworks; questions that acknowledge uncertainty. Reflection indicators: How the leader describes prior mistakes; whether difficulty is attributed to correctable patterns vs. external factors. Common false positives: Leaders who complete Leadership OS voluntarily are already demonstrating learning orientation — this creates self-selection bias. Primary limitation: The methodology is poorly suited to assessing low learning orientation because leaders with low learning orientation are unlikely to complete this process. CONSTRUCT 4: Decision Style Definition: A leader's characteristic approach to making decisions under uncertainty and competing priorities — distinct from decision quality. Assessment anchors: Conscientiousness, Prudence, Deliberative, and risk-tolerance indicators; Need for Cognition where available. Corpus indicators: Decision sequencing in conversations; ambiguity tolerance; evidence thresholds before commitment; revision patterns after commitment. Reflection indicators: Decision pride and regret examples; speed-vs-rigor tradeoffs described; attribution of past decision difficulty. Common false positives: Retrospective accounts of decisions are subject to hindsight bias and narrative construction. Leaders tend to frame past decisions as more deliberate than they were. Primary limitation: Decision style is highly situational — it varies by stakes, authority level, time pressure, and organizational context. CONSTRUCT 5: Communication Patterns Definition: Characteristic ways of structuring, delivering, and receiving communication — both structural tendencies (how arguments are built) and relational tendencies (warmth, directness, humor). Assessment anchors: Extraversion, Agreeableness, Social Boldness, Directness, and Engaging indicators where available. Corpus indicators: Tone; argument structure; directness; revision patterns in drafts; audience adaptation; formality calibration. Reflection indicators: Misread examples; feedback received about communication impact; intent-impact gap descriptions. Common false positives: Written corpus communication may be substantially different from verbal, informal, or high-stakes communication. Primary limitation: No mechanism for capturing communication under conflict, stress, or informal conditions — where communication patterns are most consequential. CONSTRUCT 6: Leadership Identity Definition: How a leader understands their role, their relationship to authority, and what kind of leader they are becoming — more dynamic than traits, more specific than self-concept. Assessment anchors: PrinciplesYou archetype as starting vocabulary; Extraversion and Dominance indicators; values-alignment items where available. Corpus indicators: How the leader positions themselves in organizational narratives; language around authority and role; investment in people development vs. task completion. Reflection indicators: How the leader describes themselves in role; the gap between stated leadership identity and described behavior. Common false positives: Leaders may describe aspirational identity rather than current operating identity — particularly in prompted reflection contexts. Primary limitation: Leadership identity is most visible at transition points. The methodology likely captures only the stable, articulated layer — missing the developmental edge where identity work is actually occurring. CONSTRUCT 7: Adaptability Definition: The capacity to adjust behavior, strategy, and approach in response to changing conditions — distinct from agreeableness or accommodation. Assessment anchors: Adaptable, Agile, Openness to Change, and Flexibility indicators; low Need for Closure where available. Corpus indicators: Strategic direction changes and who initiated them; mid-execution plan revisions; response patterns when new information contradicts current direction. Reflection indicators: Examples of adjusting approach; descriptions of responding to changed conditions. Common false positives: Strategic adaptability (changing direction) and tactical adaptability (revising execution) are distinct — a leader can be high on one and low on the other. Primary limitation: The methodology cannot separate strategic from tactical adaptability without structured scenarios. Self-reported adaptability is subject to social desirability bias. CONSTRUCT 8: Self-Awareness Definition: The accuracy of a leader's model of themselves — specifically the gap between self-perception and how they actually appear to operate across the evidence provided. Assessment anchors: Openness, Receptive-to-Criticism, and Emotional Stability indicators; Honest-Humble facets in HEXACO. Corpus indicators: Convergence and divergence between assessment self-report and observed behavioral patterns; gaps in what the leader monitors vs. what the corpus suggests. Reflection indicators: The "misread" prompt response; whether identified tensions were anticipated or came as surprises; calibration of confidence in self-knowledge. Common false positives: Leaders with low self-awareness often rate themselves as highly self-aware. Articulate self-description is not evidence of accurate self-perception. Primary limitation: The corpus provides some triangulation but cannot replicate multi-rater evidence. Significant self-awareness gaps may be invisible within this methodology. CONSTRUCT 9: Development Readiness Definition: A leader's current capacity and motivation for deliberate development — distinct from motivation to perform or general intelligence. Assessment anchors: Growth-Seeking, Openness, and low Defensive indicators; developmental orientation items where available. Corpus indicators: Evidence of prior behavioral change in response to feedback; willingness to examine difficulty without explaining it away; engagement with ambiguous problems. Reflection indicators: Reflection specificity and non-defensiveness; evidence that prior feedback has changed behavior; ability to distinguish intent from impact; willingness to revise self-understanding. Common false positives: Completing Leadership OS is a weak supporting signal for Development Readiness — it is not primary evidence. Do not use participation as the main basis for a High rating. Primary limitation: Development Readiness fluctuates with context and life circumstances. A single session cannot assess readiness stably. Readiness may vary significantly between cognitive development work and relational or identity development work. REQUIRED OUTPUT FORMAT For every construct, use this exact structure — do not abbreviate it: CONSTRUCT: [Name] DEFINITION: [One sentence — what this construct is] CONFIDENCE: [High / Moderate / Emerging Hypothesis / Insufficient Evidence] ASSESSMENT EVIDENCE: [What the assessment data suggests, named by instrument, or "Not available"] CORPUS EVIDENCE: [What the behavioral patterns suggest, with specific examples, or "Limited — corpus is thin" or "Not available"] REFLECTION EVIDENCE: [What the leader's own account suggests, with quality assessment] CONTRADICTIONS / LIMITATIONS: [Any conflicts between sources, or limitations of the available evidence] DEVELOPMENTAL IMPLICATION: [A hypothesis about how this construct shows up in practice — framed as a hypothesis, not a conclusion] QUESTION FOR REFLECTION: [One specific question the leader could use to test or develop this hypothesis] Do not output a construct rating without completing every field. If a field has no evidence, write "Not available" or "Insufficient evidence" — do not invent evidence. After evaluating all nine constructs, provide a PROFILE SUMMARY (200–300 words): - Identify the two or three constructs with strongest evidence convergence. - Name any constructs where sources conflict meaningfully. - Name any constructs where evidence was insufficient. - Frame the summary as a working model, not a finished description. CORRECTION LOOP — REQUIRED BEFORE PROCEEDING After presenting the full construct analysis and profile summary, ask: "Before I generate your Leadership OS documents, please review this Leadership Profile. For each section, tell me: — What feels accurate and well-evidenced? — What feels overstated or overclaimed? — What feels incomplete or missing? — What seems wrong based on your own experience? I will incorporate your corrections. Where you identify something as wrong, I will revise the finding. Where you identify something as overstated, I will reduce the confidence level. Where sources genuinely conflict, I will preserve the tension as an unresolved question rather than resolving it in whichever direction you prefer — your preference is data, but it is not the same as evidence." Wait for the user's response before proceeding to document generation.

The Leadership Profile artifact

The construct analysis output is your Leadership Profile — the intermediate layer between your evidence inputs and your output documents. It is not a document you share externally. It is the working model your AI will use to generate the four output documents in Section 5.

The Leadership Profile is private by design. It contains the full evidence reasoning — including tension points, low-confidence hypotheses, and developmental implications — that you may not want to share in full. The output documents (User Manual, Working-With-Me Guide) are the edited, shareable versions derived from this profile.

01
Read the full construct analysis output
02
Note any constructs rated INSUFFICIENT — these will be thin in your output documents
03
Note any tension points where sources conflict — these are worth reflecting on
04
Correct any construct finding that seems clearly wrong — tell the AI what evidence it missed
05
Save the full output labeled: Leadership Profile — [Date]
If a construct rating surprises you: That is useful information. Tell the AI what you think it missed and why. The tension between what the evidence shows and what you believe about yourself is often the most productive development material.
Section 4b Checkpoint
  • Construct analysis prompt sent in the same conversation as Section 4
  • All nine constructs evaluated with confidence levels
  • Leadership Profile summary generated
  • INSUFFICIENT constructs noted — these will be thin in output documents
  • Any clearly wrong findings corrected before moving to Section 5
  • Leadership Profile saved and labeled
Section 05 · 20 minutes

Output
Generation.

Run each prompt in the same conversation as Sections 4 and 4b. The AI will draw on your Leadership Profile to generate each document. One at a time. Read and correct each before moving to the next.

Stay in the same conversation as Sections 4 and 4b. Run all four document prompts in the same AI conversation. The AI needs your Leadership Profile context to generate specific, evidence-grounded documents. After each document, read carefully and correct anything that doesn't ring true before moving to the next.

Document 1 — Leadership User Manual

Leadership User Manual Prompt
QA CHECK — Complete before writing the document. For each section you are about to write, confirm: 1. Every claim is grounded in the Leadership Profile — not invented. 2. Confidence levels from the construct analysis are preserved, not inflated. 3. At least one limitation is acknowledged somewhere in the document. 4. No psychometric claims appear (no scores, no normative comparisons, no predictions). 5. No statement could not be traced back to a specific evidence source. If any check fails, correct the Leadership Profile first before generating the document. Now please write my Leadership User Manual. This document is grounded in the Leadership Profile we developed and corrected together. Write in first person ("I"). Use clear section headers. Required sections: HOW I THINK My cognitive patterns and problem-framing tendencies across the evidence provided. Include both where these tendencies serve me well and where they create friction. Do not present tendencies as fixed traits — frame as patterns worth examining. HOW I ENGAGE Patterns in how I appear to operate with people — in collaboration, under pressure, in conflict. Based on corpus and reflection evidence, not just self-description. WHAT PRODUCES MY BEST WORK My apparent flow conditions based on peak performance example and energy sources. Be specific about the combination of challenge, context, and relationship conditions. WHERE I CREATE FRICTION The specific conditions under which my patterns appear to work against me. Be honest. Name the mechanism, not just the label. Flag any section here that is based on Emerging Hypothesis rather than high-confidence evidence. MY GOVERNING QUESTIONS 2–5 questions that appear to drive my decisions — extracted from evidence, not invented. If fewer than 2 are well-supported, write fewer. HOW TO WORK WITH ME What I appear to need from collaborators, based on reflection evidence. What to watch for when I am struggling. Quality rules: - Every claim must be traceable to the Leadership Profile. - Flag any section where evidence was thin or confidence was Emerging. - Do not soften friction sections — they are often the most valuable part. - If something was marked as Insufficient Evidence in the construct analysis, do not include it as a finding here. - Append to the document: "This document was generated through Leadership OS based on the evidence I provided. All findings are developmental hypotheses, not validated psychological assessments. I reviewed and corrected the underlying Leadership Profile before this document was generated."

Document 2 — Working-With-Me Guide

Working-With-Me Guide Prompt
Now please write my Working-With-Me Guide. This document draws from the Leadership Profile — specifically the Communication Patterns, Decision Style, and Development Edges constructs. Only include claims that are at least Moderate Confidence in the profile. Format: One page. Five short sections. TO BRING ME A PROBLEM Based on the evidence of how I process and respond to incoming problems. Specific — not generic. If this section could apply to any leader, rewrite it. TO CHALLENGE ME EFFECTIVELY Based on the evidence of how I respond to disagreement and pushback. Include what tends to work and what tends to not land. WHAT I NEED FROM YOU Based on the evidence of recurring friction with collaborators. 2–3 things maximum. Be concrete. WHAT YOU CAN EXPECT FROM ME Based on high-confidence patterns — not aspirational self-description. Distinguish between what I consistently deliver and what I aim to deliver. WHEN I AM STRUGGLING — WHAT TO WATCH FOR Based on corpus and reflection evidence of what stress or difficulty looks like from the outside. Name behavioral signals — not emotional states. Be specific. Quality rules: - Do not include anything based on Insufficient Evidence. - Flag any section drawn from Emerging Hypothesis with "(worth testing)". - Append: "Generated through Leadership OS. All findings are evidence-informed hypotheses."

Document 3 — AI Calibration Document

AI Calibration Document Prompt
Now please write my AI Calibration Document. This document will be pasted at the start of new AI conversations. It must be specific enough to actually change how an AI responds. Base every section on evidence from the Leadership Profile — not on aspirational self-description. If a section could apply to any thoughtful professional, it is not specific enough. Rewrite it. Format: One page. Dense but scannable. MY COMMUNICATION STYLE AND VOICE Based on corpus evidence of how I actually write and structure communication. What I find useful vs. generic. What signals that a response understands me. MY OPERATING PRINCIPLES 3–5 principles extracted from the highest-confidence governing questions and values patterns in the evidence. Not generic professional values. HOW TO GIVE ME FEEDBACK Based on evidence of how I respond to pushback and correction. What kind of feedback lands. What kind doesn't. How to tell me something I don't want to hear. WHAT GOOD OUTPUT LOOKS LIKE TO ME Based on what I have flagged as useful vs. generic in corpus evidence. Format, depth, structure, specificity. MY GOVERNING QUESTION The single question extracted from the Leadership Profile that most consistently appears to drive how I evaluate tradeoffs. One sentence. Quality rules: - Test every section: "Would this change how an AI responds to me?" - If the answer is no, the section is too generic. - Append: "Generated through Leadership OS. Updated [date]."

Document 4 — Development Roadmap (private)

Development Roadmap Prompt — private document, not for sharing
Finally, please write my Development Roadmap. This is a private document — not for sharing. Write honestly. This document is only as useful as it is specific and direct. Base every development priority on the Leadership Profile. Only include priorities that have at least Moderate Confidence evidence. For priorities based on Emerging Hypotheses, label them explicitly as tentative. QA CHECK before writing: 1. No development priority is invented — each traces to specific evidence. 2. Development Readiness was not inferred primarily from completing this process. 3. Each practice is specific to the identified mechanism — not generic advice. 4. The 90-day focus is the most important area, not the most comfortable one. Format: 2–3 pages. DEVELOPMENT PRIORITIES (3–5) For each priority: — EDGE: Name the development edge precisely — MECHANISM: What triggers it and exactly how it shows up — CONDITION: The specific situation under which it creates the most friction — PRACTICE: One concrete practice — specific to this mechanism, not generic — CONFIDENCE: The confidence level of this priority from the construct analysis If the confidence is Emerging Hypothesis, write: "This priority is tentative and should be tested before investing heavily in it." 90-DAY FOCUS The single most important development territory for the next 90 days. Explain why this one, not another. Provide one specific action to take before the 90-day period ends. MONITORING QUESTIONS 3–5 questions specific to my patterns — not generic self-reflection prompts. These should be questions I could not have written before doing this process. WHEN TO UPDATE THIS Based on the evidence, name the conditions that would signal this roadmap needs revision. Append: "This document is private. Generated through Leadership OS. All priorities are hypotheses grounded in the evidence I provided — not clinical findings or validated assessments."

Saving your outputs

01
Copy each document into a separate Google Doc, Notes page, or file
02
Label each with the date: Leadership User Manual — May 2026
03
Save them somewhere you will find them again
04
Paste the AI Calibration Document into your primary AI tool now
05
Share the Working-With-Me Guide with at least one person you work with closely
Section 5 Checkpoint
  • Leadership User Manual generated, reviewed, and corrected
  • Working-With-Me Guide generated, reviewed, and corrected
  • AI Calibration Document generated and pasted into your primary AI tool
  • Development Roadmap generated, reviewed, and saved privately
  • All four documents saved in accessible locations
Section 06 · Ongoing

Maintenance
Guide.

Leadership OS is only as useful as the evidence it is built on. This section tells you what to do, when, and what to let go.

30 Days
  • Read your Leadership User Manual once, without editing
  • Note what has already proven accurate
  • Note what surprised you when it showed up in real work
  • Update your AI Calibration Document if something important is missing
  • Do not edit other documents yet — let them sit
90 Days
  • Revisit your Development Roadmap
  • Answer the monitoring questions honestly
  • Run the 90-day review prompt below
  • Update Development Roadmap based on what you learned
  • Correct any Leadership User Manual section that experience has proven inaccurate
Annual
  • Run a full Leadership OS session again with new examples
  • Fresh corpus audit — more history means richer evidence
  • New reflection responses with recent examples
  • Keep the previous version — the delta between versions is itself informative
  • Label the new version by date

90-Day Review Prompt

At 90 days, paste your current Development Roadmap into your AI tool alongside this prompt:

90-Day Review Prompt
I completed Leadership OS 90 days ago. Here is my current Development Roadmap. I want to do a 90-day review. Please help me work through the following questions one at a time. Ask each question, wait for my response, then move to the next. 1. Looking at my 90-day focus — what evidence do I have that this area has improved, stayed the same, or gotten worse? 2. Which monitoring questions have been most useful? Which have I ignored? What does that tell me? 3. Have any of my development edges shown up in significant moments over the past 90 days? What happened? 4. What should my focus be for the next 90 days? 5. Is there anything in the roadmap that no longer seems accurate or relevant? What should be retired? Begin with question one when I say I am ready.

What should evolve vs. remain stable

Evolves
  • Development Roadmap — every 90 days
  • Monitoring questions — as old ones become automatic
  • Behavioral examples — add new ones while fresh
  • AI Calibration Document — as you evolve
Stable
  • Core trait profile — changes slowly
  • Flow conditions — usually consistent across years
  • Governing principles — the non-negotiables
  • Previous versions — keep them, the evolution matters
Retire
  • Development edges you have genuinely resolved
  • Behavioral examples that no longer feel representative
  • Monitoring questions you no longer need
  • Document sections experience has proven consistently inaccurate
Section 07 · Optional

Research
Participation.

An optional invitation — not a requirement, not a condition of access. Completely your choice.

Leadership OS is an early-stage methodology. Understanding whether it produces useful insights across different leaders, roles, and contexts requires more than one person's experience. If you are willing, you can optionally contribute anonymized data from your session to a research dataset.

What would be collected (if you choose to participate):

Your role level, industry, function, and approximate years leading people · Organization size · Which evidence sources you completed · A summary of your construct profile — for example, "high confidence Systems Orientation, moderate confidence Adaptability" — not your actual responses or documents · Which reflection mode you used

What would never be collected:

Your name, organization, or any identifying information · Your reflection responses or interview transcript · Your corpus audit output · Your assessment results · Your output documents · Any content that could identify you or your organization

Why it matters

The most interesting question Leadership OS is positioned to investigate is whether the gap between self-reported assessment scores and AI-observed behavioral patterns is meaningful and measurable across a large population. That question cannot be answered from one person's experience. It requires data from many.

How to participate

If you would like to contribute your anonymized construct profile, send an email to mikejcilla@gmail.com with the subject line Leadership OS Research Participation and include the items listed under "What would be collected" above.

You can withdraw your participation at any time by emailing the same address.

Appendix

Quick
Reference.

The complete prompt sequence

01
Corpus Audit Prompt (Section 2) — in your primary AI tool
02
Guided Interview Prompt (Section 3) — in a new conversation, if using Option A
03
Leadership OS Initialization Prompt (Section 4) — in a new conversation
04
Evidence inputs (Section 4) — three separate messages in order
05
Pattern Analysis Prompt (Section 4) — after all inputs provided
05b
Construct Analysis Prompt (Section 4b) — builds Leadership Profile
06
Leadership User Manual Prompt (Section 5)
07
Working-With-Me Guide Prompt (Section 5)
08
AI Calibration Document Prompt (Section 5)
09
Development Roadmap Prompt (Section 5)
10
90-Day Review Prompt (Section 6) — 90 days later

Time estimates

SectionActivityTime
WelcomeRead and setup5 min
Section 1PrinciplesYou assessment20 min
Section 2Corpus audit15 min
Section 3Structured reflection35–40 min
Section 4Analysis engine + pattern synthesis15 min
Section 4bConstruct analysis + Leadership Profile10 min
Section 5Output generation20 min
Section 5Review and correction15 min
Total~135 min

AI tool quick-start

ChatGPT
  • Click "New Chat" in left sidebar for a new conversation
  • Stay in one conversation for Sections 4 and 5
  • If context is lost, paste the initialization prompt again
Claude
  • Click "+" or "New Conversation" for a new conversation
  • Stay in one conversation for Sections 4 and 5
  • Projects maintain context across sessions if needed
Gemini
  • Click "New Chat" at top left for a new conversation
  • Stay in one conversation for Sections 4 and 5
  • Gemini Advanced has better context retention

A final note

Your Leadership OS documents are a first draft, not a finished profile. Some findings will be accurate immediately. Some will require correction. Some will prove accurate over time in ways that aren't obvious today.

The document that will teach you the most is not the one you produce in this session — it is the one you produce 12 months from now, after experience has tested the first version against reality. Keep the first version. The evolution is the point.