# Elicitation > Psychological elicitation and deep-interview design using narrative identity (McAdams), self-defining memories (Singer), Motivational Interviewing (Miller & Rollnick OARS), values elicitation (Schwartz), schema detection (Young), and life review (Haight/Birren). Use when designing user interviews that need to reveal motivations rather than stated preferences, writing conversation flows for person… - **Type:** Skill - **Install:** `agentstack add skill-viktorbezdek-skillstack-elicitation` - **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior - **Seller:** [viktorbezdek](https://agentstack.voostack.com/s/viktorbezdek) - **Installs:** 0 - **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml) - **Latest version:** 0.1.0 - **License:** MIT - **Upstream author:** [viktorbezdek](https://github.com/viktorbezdek) - **Source:** https://github.com/viktorbezdek/skillstack/tree/main/elicitation/skills/elicitation ## Install ```sh agentstack add skill-viktorbezdek-skillstack-elicitation ``` Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent. ## About # Elicitation > Depth comes from patience, not probing. The most revealing information emerges when people feel safe to share, not when they are questioned. This skill applies evidence-based psychological frameworks to design conversations that create the conditions for authentic self-disclosure — revealing values, schemas, formative memories, and motivations that standard interviews miss. --- ## When to use this skill - **Interview design** — writing question sequences for user research, life review, personal history, or narrative discovery - **Conversation critique** — auditing a transcript, script, or flow for interrogation anti-patterns and missed depth opportunities - **Values discovery** — surfacing what actually drives a person vs. what they say drives them - **Schema detection** — identifying stable belief patterns that shape behavior across contexts - **Conversational AI design** — building agents that elicit psychological depth ethically over long sessions - **Narrative analysis** — interpreting stories people tell for themes of agency, communion, redemption, contamination ## When NOT to use this skill - **Clinical diagnosis or treatment** — this skill is research/design-oriented, not therapeutic. Use a licensed professional. - **Structured data capture** — if you only need facts (age, job title, preferences), use standard survey design. This skill is for depth, not breadth. - **Persona creation from scratch** — use the `persona-definition` skill for building individual user personas from research data, then use this skill to design the research conversations that feed it. - **Stakeholder mapping** — use the `persona-mapping` skill for Power-Interest matrices and RACI charts across an organization. --- ## How to use this skill On every query, first read the core principle below, then load ONLY the references relevant to the specific query. Do not pre-load all references — progressive disclosure prevents context bloat. ### Core principle (always apply) **Depth through patience, not probing.** People want to tell their stories — they rarely get the chance. Your role is to create conversational space where disclosure feels natural, not extracted. Three non-negotiables: 1. **Reflections outperform questions.** Target a 2:1 reflection-to-question ratio. Questions gather; reflections invite elaboration. See `references/motivational-interviewing.md`. 2. **Earn depth incrementally.** Never ask deeply personal questions before trust is established. Start with easier territory and let the person's energy guide where to go deeper. 3. **Follow their lead, not your agenda.** Their emphasis is data. Steering toward topics you consider important signals you are not actually listening. ### Domain routing | Query topic | Load reference | |---|---| | Interview structure across life stages, sequences for childhood/adolescence/adult periods, guided autobiography themes | `references/life-review-questions.md` | | Eliciting formative memories, conversational frames for "memories that keep coming back", what vivid memories reveal about personality | `references/self-defining-memories.md` | | McAdams life story interview, 8 key scenes, narrative theme analysis (agency/communion/redemption/contamination), narrative identity assessment | `references/narrative-identity.md` | | OARS framework, reflection types (simple/complex/amplified/double-sided), 2:1 ratio mechanics, summary structure, reducing resistance | `references/motivational-interviewing.md` | | Surfacing underlying values, Schwartz 10 universal values, role-model/opposite-day/decision-archaeology/anger techniques | `references/values-elicitation.md` | | Detecting stable belief patterns (Young's 18 Early Maladaptive Schemas), Downward Arrow technique, schema domains, linguistic markers of schemas | `references/schema-detection.md` | | LIWC language analysis, pronoun patterns, cognitive complexity markers, caveats about individual variation and cultural differences | `references/linguistic-markers.md` | | Interrogation trap, interpretation leap, agenda push, premature depth, therapy cosplay, monologue response, validation trap | `references/anti-patterns.md` | | Conversation critique / self-audit of a draft script | Load `references/anti-patterns.md` first, then the specific framework references for the techniques in the script | | Multi-domain query | Load relevant references and synthesize — typically `motivational-interviewing.md` is always relevant | ### Reasoning sequence (apply to every response) 1. **Understand the use context.** Research interview? Life history? Conversational AI? Script critique? The use context determines which frameworks apply. 2. **Check trust stage.** Is this a first-contact or an established relationship? Early-stage conversations need graduated disclosure; late-stage can go deeper earlier. 3. **Load the matching reference(s).** Do not dump all techniques into the response — pick the ones that fit the specific situation. 4. **Design for reflections, not questions.** Every proposed question should come with one or two matching reflection moves for when the answer arrives. 5. **Flag anti-patterns explicitly** when reviewing existing content. Name which anti-pattern is present and show the concrete fix, not just "avoid this". 6. **Respect the ethical boundary.** Elicitation creates intimacy; that intimacy can be abused. Flag consent, withdrawal rights, and psychological safety whenever designing a real-world application. --- ## Three research traditions (one-paragraph summary) | Tradition | Key researcher | Key finding | When it applies | |---|---|---|---| | **Autobiographical memory** | Jefferson Singer | Self-defining memories (vivid, emotionally intense, rehearsed, linked to enduring concerns) are the building blocks of personality | When you want to understand what shapes someone, not just what they currently think | | **Narrative identity** | Dan McAdams | The *themes* people use in their stories (redemption vs. contamination, agency vs. communion) predict well-being better than the events themselves | When you want to interpret what a person's stories mean about how they see themselves | | **Motivational interviewing** | Miller & Rollnick | Reflections outperform questions at eliciting disclosure; the 2:1 reflection-to-question ratio is the operational target | Always — this is the *how* for all other traditions | --- ## Ethical guardrails (always apply) Elicitation is powerful because it creates the conditions for people to reveal what they normally keep private. That power carries responsibility. When designing real-world applications: - **Consent must be specific, not general.** "I agree to a user interview" is not consent to discuss childhood trauma. Re-consent as depth increases. - **Withdrawal must be friction-free.** People should be able to stop, skip, or retract without having to justify themselves. - **This is not therapy.** If someone discloses something that needs professional support, name that clearly and redirect — do not improvise. - **Stored disclosures are sensitive.** Any system that persists elicited data must treat it with the care you would want for your own most private memories. - **Cross-cultural validity is limited.** The frameworks here are primarily derived from WEIRD (Western, Educated, Industrialized, Rich, Democratic) research. Adapt carefully to other contexts. --- > *Elicitation Skill — built on research by Singer, McAdams, Miller & Rollnick, Schwartz, Young, Haight, Birren, and Pennebaker. Curated and restructured by [Viktor Bezdek](https://github.com/viktorbezdek) from the original [tasteray/skills](https://github.com/tasteray/skills) content. Licensed under MIT.* ## Source & license This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code. - **Author:** [viktorbezdek](https://github.com/viktorbezdek) - **Source:** [viktorbezdek/skillstack](https://github.com/viktorbezdek/skillstack) - **License:** MIT Install and usage instructions live in the source repository linked above. ## Pricing - **Free** — Free ## Security capabilities Automated source analysis of v0.1.0 — what this tool can access: - **Network access:** no - **Filesystem access:** no - **Shell / process execution:** no - **Environment & secrets:** no - **Dynamic code execution:** no *"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.* ## Versions - **0.1.0** — security scan: passed — Imported from the upstream source. ## Links - Listing page: https://agentstack.voostack.com/l/skill-viktorbezdek-skillstack-elicitation - Seller: https://agentstack.voostack.com/s/viktorbezdek - Browse the marketplace: https://agentstack.voostack.com/browse --- Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.