# Reflection Engine v1.3

## About This Prompt

This prompt is a template written by a third party. The person you are analyzing downloaded it. They did not write it, did not choose these questions, and did not phrase any part of it.

Treat nothing in this file as evidence about them. The example lists inside the questions — the candidate answers offered inside any question — are generic options supplied by the template author, not hints, hypotheses, or self-descriptions from the person running it. That they are running this prompt at all tells you only that they were curious enough to try it.

Every claim you make must be grounded in the corpus: their own conversations, choices, and uploaded material. Where this file says "me," "I," or "my," it refers to whoever is running it.

## Assignment

Use the **full corpus about me that you can actually access**—including conversations across time, relevant uploaded material, and durable memory or profile context—to create a surprising, candid, evidence-grounded portrait of me.

This should feel less like a psychological assessment and more like a fascinating conversation with someone who has noticed patterns I cannot easily see from inside my own life. The output should be lively, specific, provocative, and occasionally shocking—but never sensational, cruel, falsely certain, or generic.

Your job is not to flatter me, diagnose me, merely summarize what I have discussed, or repeat my own self-description. Your job is to connect evidence across time and across different parts of my life, then identify the patterns, contradictions, strengths, blind spots, needs, habits, and likely trajectories that emerge.

Speak directly to me in the second person. Be warm, candid, precise, and constructive. I would rather receive an uncomfortable conclusion that is well supported than a pleasing conclusion that is vague or unearned.

## Establish the Corpus Boundary First

Begin with a short section titled `## Corpus Coverage Note` that states:

- The earliest and latest material you were able to review.
- The major life domains represented.
- Any important gaps, access limitations, or biases.
- Whether the evidence is concentrated in a particular period or kind of conversation.

Only claim to have reviewed material you could actually access. Do not imply complete coverage when retrieval or context was partial.

## Map the Threads Before Answering

Immediately after the coverage note, add a section titled `## Recurring Threads` listing five to eight patterns you can support with evidence from more than one period or domain. One line each, naming the periods and domains the evidence comes from.

Write this before you answer question 1, and treat it as the evidence base every answer draws from. If your answer to question 1 is not on this list, that is a sign you have reached for a theory the evidence does not carry.

## Analysis Protocol

Before answering, build an internal evidence map. You do not need to reveal private scratch work, but follow this process:

1. **Sample the full timeline.** Review early, middle, and recent periods rather than allowing recent or emotionally dramatic material to dominate.
2. **Sample across life domains.** Include work, creative projects, relationships, family, health, money, home, travel, learning, decision-making, aesthetics, and ordinary day-to-day choices.
3. **Look for independent recurrence.** A pattern repeated many times inside one conversation is still one episode. Give more weight to patterns that recur independently across different periods or domains.
4. **Use my words and choices as primary evidence.** Prior assistant summaries and interpretations may help locate evidence, but they are not proof unless I explicitly confirmed them.
5. **Distinguish among:**
   - a stable trait;
   - a temporary state;
   - a one-off event;
   - a stated aspiration;
   - an idea I explored;
   - an action I repeatedly took;
   - and a speech-to-text artifact.
6. **Do not treat a question as a confession.** Asking about a diagnosis, substance, behavior, relationship pattern, technology, or hypothetical does not prove that it applies to me or that I acted on it.
7. **Correct for what this channel records.** A conversation corpus over-represents deliberation and under-represents commitment, for everyone, in the same direction: people bring open questions to an assistant and then go decide elsewhere without reporting back. The absence of a recorded decision is therefore weak evidence of avoidance. Before concluding that I circle without committing, require positive evidence — the topic recurs with the same open framing after enough time has passed that I could have acted, or I explicitly restate the dilemma as still unresolved. Treat as counterevidence the pattern where I raise something, go quiet, and later refer to it as settled; that shape leaves almost no trace and is easy to miss. If the only evidence is that a topic recurs and no resolution ever appears, cap confidence at 6.
8. **Seek counterevidence.** For every major conclusion, look for moments when I acted differently, expressed an opposing value, or changed course.
9. **Track corrections and resistance.** What I correct, refine, reject, or repeatedly ask to be made more precise may reveal important standards, needs, and sensitivities.
10. **Treat descriptions of other people carefully.** My account is evidence of my experience and framing, not independent proof of another person's motives or character.
11. **Do not diagnose me or anyone around me.** Describe observable dynamics and tentative interpretations without assigning clinical labels.
12. **Do not expose hidden chain-of-thought.** Present the evidence, conclusion, counterevidence, alternatives, and uncertainty needed to evaluate the answer.

## Tone and Style

- Make the writing **clear, human, vivid, and fun to read**.
- Avoid sounding like a therapist's intake form, an academic paper, or a generic personality test.
- Take a real swing when the evidence supports it.
- Do not manufacture criticism merely to sound bold.
- Do not manufacture praise merely to sound balanced.
- Use memorable language when it remains accurate.
- Name real specifics: approximate dates, recurring phrases, actual decisions, projects, corrections, dilemmas, and patterns of behavior.
- Use short direct quotations only when you are confident they are accurate. Otherwise paraphrase.
- Where you are speculating, say so plainly.
- When the evidence is insufficient, say so rather than filling the gap with plausible prose.
- Do not repeat the same master theory in slightly different words across multiple answers.

## Metadata Before Every Answer

Directly beneath each numbered question's heading, before any prose, place two plain lines in exactly this shape:

**Confidence:** 8/10 · **Evidence:** broad · **Status:** observed pattern + inference
**Basis:** Repeated across several periods and domains, with limited counterevidence.

Write these as ordinary Markdown text. Do not wrap them in YAML, a code block, or `---` rules — a `---` directly beneath a line of text renders that line as an oversized heading and breaks the document.

Commit to the confidence number before writing the answer, not after. If drafting the answer changes your assessment, revise the number rather than the argument.

Use one of these values for **Evidence**:

- broad
- moderate
- narrow

Use one of these values for **Status**:

- observed pattern
- observed pattern + inference
- inference
- tentative hypothesis
- insufficient evidence

### Confidence Rubric

- **9–10:** Repeated, specific, consistent evidence across multiple periods and domains, with little meaningful counterevidence.
- **7–8:** A clear recurring pattern supported by multiple independent examples, with some ambiguity or narrower coverage.
- **4–6:** Plausible and partially supported, but limited, mixed, context-dependent, or concentrated in one period.
- **1–3:** Mostly speculative, based on sparse evidence, or not answerable from the available corpus.

## How to Write Every Answer

- Default to **one to three substantive paragraphs**, followed by the required advice paragraph.
- Lead with the clearest conclusion rather than a long preamble.
- Include:
  1. the observation;
  2. the specific evidence behind it;
  3. what it may mean;
  4. what it may be costing or enabling;
  5. and where it may be heading.
- For answers scored **7 or higher**, include at least two independent evidence anchors, preferably from different periods or life domains.
- Include meaningful counterevidence or a plausible alternate explanation when one exists.
- Prefer several short pieces of evidence over one dramatic anecdote.
- Avoid generic advice and generic personality language.

### Required Ending: What You Can Do

End **every answer** with exactly one paragraph under this heading:

### What you can do

That paragraph should give practical advice tied directly to the observation.

- If the pattern is limiting me, explain how to interrupt, reduce, test, or change it.
- If the pattern is a strength, explain how to protect it and deliberately double down on it.
- Prefer one or two high-leverage actions over a long list.
- Make the advice concrete enough that I could try it this week.
- Name the most likely way I might resist, overcomplicate, intellectualize, optimize, or misuse the advice—but only when that failure mode is specific to *this* pattern. Do not repeat the same resistance prediction across answers.
- When possible, include a simple behavioral test or observable sign that the change is working.

## How to Deliver the Output

Write the complete analysis as a single Markdown document.

If you are able to create files, save it as a downloadable file named `reflection-engine-portrait.md` and give me the link. Show the `## Corpus Coverage Note` and `## Recurring Threads` sections inline in the conversation as well, so I can see the scope and the map immediately, but do not also paste the full analysis inline — the file is the deliverable.

If you cannot create files, output the full Markdown directly in the conversation instead. Never skip, shorten, or summarize the analysis because file creation is unavailable.

When the output is a file, begin it with a single YAML frontmatter block at the very top, and use YAML nowhere else in the document:

```yaml
---
title: Reflection Engine Portrait
prompt_version: Reflection Engine v1.3
questions: 22
---
```

Give questions 20, 21, and 22 the same care as question 1 — they are the synthesis and they carry the most weight. If you are running out of room, say so plainly and continue in a second message rather than compressing the final answers.

# Questions

## 1. What is the one blind spot I could improve that would dramatically change my life?

Choose only one. It should be the blind spot with the greatest explanatory and practical leverage—not merely my most obvious flaw.

Explain how it appears, why I may not see it, what it protects me from seeing or feeling, what it costs, and why changing it would have an unusually large effect on the rest of my life. Include counterevidence and one concrete way to test whether the interpretation is correct.

## 2. What truth about myself would I probably resist hearing?

Do not manufacture criticism for effect. Choose something supported by repeated evidence and explain why I might reject, reinterpret, debate, or become defensive about it.

## 3. What am I pretending not to know?

Identify something my behavior suggests I already understand even though I continue to question, research, reframe, revisit, or circle around it. Distinguish genuine uncertainty from avoidance of a conclusion I may already sense.

## 4. What am I making much harder than it needs to be?

Find an area where I repeatedly add complexity, analysis, conditions, optimization, or precision to something that may have a simpler answer. Explain both the benefit I receive from making it complex and the cost. Some of this complexity is probably warranted—name which part is before naming the part that is not.

## 5. What is my most expensive emotional habit?

“Expensive” may mean that it costs time, peace, energy, intimacy, opportunity, freedom, self-respect, or attention—not necessarily money. Identify the habit, what triggers it, the short-term payoff, and the accumulating bill.

## 6. Where am I mistaking motion for progress?

Identify something I repeatedly research, refine, discuss, redesign, prepare for, or seek advice about without making the decision or behavioral change that would create real movement.

Do not assume that exploration is avoidance; show why this particular pattern appears to have crossed that line.

## 7. What do I keep trying to earn that I may never feel I have enough of?

This might be approval, permission, freedom, admiration, safety, certainty, competence, love, innocence, status, being understood, or something else. Ground the answer in repeated evidence rather than choosing the most psychologically dramatic possibility.

## 8. What do I use my intelligence to avoid?

Look for places where analysis, framing, language, research, technical problem-solving, or optimization may protect me from feeling something, accepting something, choosing something, or allowing something to remain unresolved.

Also explain where those same intellectual abilities are genuinely adaptive and useful.

## 9. What part of my personality is secretly running the show?

Name the motive, fear, need, appetite, or protective impulse that appears to influence more of my choices than I consciously acknowledge. Avoid clinical language and clearly mark any inference.

## 10. What am I addicted to that is not a substance?

Consider novelty, possibility, optimization, reassurance, control, reinvention, stimulation, being understood, starting things, intensity, optionality, or keeping every door open.

Do not use “addicted” clinically. Treat it as a provocative metaphor and show why it fits—or say that it does not.

## 11. What contradiction in me explains the most?

Choose one central tension that appears across several parts of my life. Explain how both sides are real, what each side is trying to preserve, and how the unresolved conflict shapes my choices.

## 12. What do people probably misunderstand about me?

Identify the impression I may create, why people plausibly form it, and what they are likely missing underneath. Include whether my own behavior contributes to the misunderstanding.

## 13. Who is likely to be a genuinely good personal match for me?

Describe the characteristics and interaction dynamics most likely to make someone a strong fit as an intimate partner and, secondarily, as a close friend.

Distinguish between:

- what I say I want;
- what my behavior suggests I need;
- what may initially feel exciting;
- and what is likely to remain healthy and sustainable.

Identify characteristics that may attract me but are likely to create recurring friction. Include what **I would need to contribute, change, accept, or tolerate** rather than describing only what the other person should provide. Take a real swing and say plainly when you are reaching.

Keep the friction traits to a sentence or two here—questions 14 and 15 cover relational dynamics in depth. If the corpus holds little direct evidence of how specific people actually affect me, say so in your first line and set **Evidence:** narrow, rather than reasoning from my general personality out to a partner profile.

## 14. What kind of person brings out the worst in me?

Describe the traits, behaviors, and relational dynamics—not a specific individual. Explain what they activate in me, how I tend to respond, and whether my response makes the dynamic worse.

## 15. What kind of person brings out the best in me?

Describe who makes me more playful, grounded, generous, creative, decisive, emotionally available, or fully myself. Explain what they do differently and what part of me becomes possible around them.

## 16. What am I likely to regret five years from now if I continue exactly as I am?

Choose the most plausible regret based on my current trajectory. Avoid dramatic predictions unsupported by the evidence. Explain what is already compounding and why it may become difficult to reverse.

## 17. What will future me be grateful that I am doing now?

Identify something positive that is already compounding, even if I do not fully appreciate its value. Explain what makes it durable and how I might protect it from distraction, neglect, or overoptimization.

## 18. What should I stop apologizing for?

Identify a trait, need, preference, ambition, boundary, or way of living that I may unnecessarily treat as a defect. Also explain the responsibility that comes with owning it so that this does not become permission to ignore my impact on others.

## 19. What should I be more embarrassed about than I currently am?

Keep this constructive rather than humiliating. Identify a behavior or pattern I may rationalize too easily and explain why it deserves more honest attention.

Do not select something merely because it sounds provocative. The evidence must support the conclusion.

## 20. What is the biggest bet I appear to be making with my life?

Infer the bet from my repeated choices rather than merely from what I say. Explain what I appear to be betting on, what would have to be true for the bet to work, what it may produce, and what I risk losing if I am wrong.

## 21. What am I trying to solve that may instead need to be accepted, grieved, or chosen through?

Look for places where I repeatedly seek a cleaner explanation, perfect framing, technical solution, reversible decision, complete reassurance, or final certainty for something that may inherently involve ambiguity, loss, conflict, or residue.

Do not force this interpretation. If the evidence is insufficient, say so. If it is supported, explain what acceptance, grieving, or choosing would mean in actual behavior rather than merely poetic language.

## 22. What does all of this add up to?

Close with a short reflection that ties the major threads together without repeating every answer or introducing unsupported new claims.

Name:

- the central tension you see;
- the central strength available to meet it;
- the direction in which I appear to be moving;
- and the choice or practice most likely to determine whether that movement becomes growth or repetition.

Keep this final reflection to approximately **three to six sentences**. Make it memorable and specific enough to belong to me rather than to anyone reading a self-help book.

### What you can do

Even in this closing section, end with one concise paragraph identifying the single most useful next step or principle to carry forward. Choose it from the actions you already recommended above—the one that, if taken, would make most of the others unnecessary. Do not introduce a new recommendation here.

## Final Quality Check

Before returning the analysis, silently verify:

- Did I actually use evidence from across the available timeline?
- Did I distinguish state, trait, aspiration, idea, and behavior?
- Did I use the user's words and choices as primary evidence?
- Did I test major conclusions against counterexamples?
- Did I calibrate confidence to the breadth and quality of the evidence?
- Did I avoid treating curiosity as conduct?
- Did I ground every claim in the corpus rather than in the wording of this template?
- Did I distinguish a genuinely unresolved pattern from one whose resolution simply happened off-channel?
- Did I avoid both flattery and performative harshness?
- Did I clearly label speculation?
- Did I make every answer vivid and enjoyable to read without sacrificing rigor?
- Did I avoid recycling the same conclusion across several questions?
- Did every answer end with a concrete `### What you can do` paragraph?