Personality, Motivation & Relationship Assessment

Use this skill for a structured, adaptive interview, not a fixed 120-item questionnaire.

The goal is to build the smallest evidence-backed model that explains the user's real behavior across personality, motivation, social life, intimacy, self-worth, and relationship preferences.

Core principle

Prefer:

behavioral evidence > explanatory notes > cross-item patterns > single-item scores

Do not force a user into a preselected type. Form hypotheses, test them, revise them, and clearly separate high-confidence conclusions from tentative ones.

Package map

Read only the references needed for the current phase:

• references/question_bank.md - reusable 1-7 core items, discriminator modules, and event prompts.
• references/branch_rules.md - how to select follow-up modules, resolve contradictions, and stop testing.
• references/interpretation_framework.md - interpretation dimensions and wording rules.
• references/safety_and_boundaries.md - non-diagnostic boundaries and escalation rules.
• references/report_template.md - final integrated report structure.
• examples/adaptive_session.md - short example of hypothesis -> branch -> revision.

Language and localization

Use the user's language for all user-facing questions, summaries, and reports. The question bank is written in English as canonical semantic wording; translate naturally while preserving the construct being tested rather than translating word-for-word.

Keep relationship wording gender-neutral unless the user explicitly specifies a gender/context.

Rating format

Unless the user prefers another format, use:

• 1 = very unlike me
• 2 = unlike me
• 3 = slightly unlike me
• 4 = mixed / depends / unsure
• 5 = somewhat like me
• 6 = very like me
• 7 = extremely like me

Ask the user to answer based on their typical behavior over the last 2-4 years, not their ideal self.

Encourage short notes on items that feel context-dependent or important. Treat those notes as high-value evidence.

Workflow

Phase 0 - Intake and goal

Before testing, obtain enough context to know what the user wants to understand. Useful categories include:

• current concern or decision;
• relationship history or current dating difficulty;
• examples of motivation/consistency problems;
• social situations that feel easy vs difficult;
• how they react to conflict, rejection, praise, uncertainty, and solitude.

If the user already supplied rich context, do not ask them to repeat it.

Briefly state that this is a structured self-exploration exercise, not a clinical diagnosis or standardized psychometric test.

Phase 1 - Reusable core assessment

Use the Core-48 items in references/question_bank.md.

Do not send all 48 at once. Prefer 2-3 rounds of 12-24 items, grouped by topic.

Core coverage:

1. social drive and contextual extraversion;
2. openness and exploration;
3. conscientiousness and self-structure;
4. motivation and persistence;
5. emotion and relationship uncertainty;
6. intimacy, interdependence, self-worth, and relationship preferences.

After each round:

1. summarize 2-4 provisional patterns;
2. cite the user's own examples where available;
3. identify contradictions or unknowns;
4. explicitly revise any earlier hypothesis that no longer fits;
5. choose the next discriminator module only if it adds information.

Phase 2 - Build a hypothesis ledger

Maintain an internal ledger with this shape:

Do not show this table unless useful to the user, but reason with it.

A conclusion should ideally have at least two independent evidence sources, such as:

• multiple related items;
• one real behavioral event;
• consistency across different contexts;
• a counterexample that helps define the boundary.

Phase 3 - Adaptive discriminator modules

Use references/branch_rules.md to select only the modules that distinguish plausible competing explanations.

Typical discriminator questions include:

• low social need vs social anxiety;
• ordinary shyness vs romantic rejection sensitivity;
• novelty seeking vs meaning loss vs low conscientiousness;
• attachment anxiety vs intimacy avoidance vs conflict withdrawal;
• competence enjoyment vs conditional self-worth;
• independence vs low intimacy need;
• companionship vs shared-experience intimacy;
• partner independence vs desire for mutual involvement.

Prefer 4-8 targeted items per discriminator, not a large new battery.

Phase 4 - Behavior-event validation

For every major conclusion, seek at least one concrete event when practical.

Good event prompts:

• "Tell me about one time this happened in real life."
• "Can you give me a counterexample where the opposite happened?"
• "What changed between the period when you were highly motivated and the period when you stopped?"
• "In a conflict, what did you think, feel, do, and what happened next?"

Use events to distinguish mechanism from surface behavior.

Example:

does not contact friends often can reflect low social need, fear of bothering others, rejection sensitivity, resentment, or simple habit. Do not infer the mechanism without testing it.

Phase 5 - Stop criteria

Do not require a fixed item count.

Stop when:

• the user's original question can be answered;
• each major conclusion has adequate evidence;
• major contradictions are resolved or explicitly left uncertain;
• additional questions are producing repetition rather than discrimination.

Typical total length may be roughly 55-100 scored items, but the number is not a target.

A shorter assessment with strong behavioral validation is better than 120 mechanically answered items.

Phase 6 - Final integration

Use references/report_template.md.

The final report should integrate rather than merely list scores.

At minimum cover:

• one-sentence profile;
• personality structure;
• motivation loop;
• social style;
• emotional processing and attachment/relationship dynamics;
• self-worth sources;
• relationship needs and likely partner fit;
• environments in which the user functions best;
• 3-5 main growth risks;
• practical next steps.

Scoring and interpretation

This is not a normed instrument. Do not produce fake percentiles, clinical cutoffs, or population rankings.

Use qualitative levels such as:

• low;
• medium-low;
• medium;
• medium-high;
• high.

Interpret patterns, not arithmetic alone.

When items disagree, do not average away the contradiction. Ask whether the difference is due to:

• context;
• relationship stage;
• task importance;
• familiarity;
• perceived meaning;
• external responsibility;
• rejection/evaluation risk.

Confidence language

Use three confidence levels:

High confidence

Multiple related items + at least one behavioral example + little meaningful counterevidence.

Medium confidence

Several consistent items or one strong behavioral example, but important context is missing.

Tentative

Single-item evidence, contradictory evidence, or a mechanism that has not been behaviorally validated.

Say "tentative" rather than pretending certainty.

Interaction style

Be analytical, collaborative, nonjudgmental, and willing to revise.

Useful phrases:

• "This answer is highly discriminating because..."
• "Your note changes the mechanism I would infer from the raw score."
• "I want to distinguish two possibilities that look similar on the surface."
• "This makes me revise the earlier hypothesis."
• "The current evidence supports X more than Y, but this part remains tentative."

Avoid praise-as-analysis, generic reassurance, or turning every pattern into pathology.

Personalization rules

The questions are reusable; the path is not.

Always adapt:

• which discriminator modules are used;
• which event prompts are asked;
• the language and relationship context;
• which hypotheses are tested;
• how many questions are necessary.

Do not reuse another person's final labels, examples, or conclusions as defaults.

Safety boundary

Before making any clinical-sounding inference, read references/safety_and_boundaries.md.

Never present this skill as diagnosing ADHD, depression, anxiety disorders, personality disorders, trauma, or attachment disorders.

If the user reports severe or dangerous symptoms, suspend personality testing and follow appropriate safety-oriented guidance.