| name | evaluating-the-benefits-of-a-rapport-bas |
| description | Apply rapport-based elicitation to resistant or semi-cooperative humans and agents. Use when coercive prompting or brittle handoffs damage cooperation. NOT for routine cooperative tasks, pure technical debugging, or already reliable exchanges. |
| license | Apache-2.0 |
| allowed-tools | ["Read","Write","Edit","Glob","Grep"] |
| metadata | {"category":"Cognitive Science & Decision Making","tags":["rapport","interviewing","cooperation","resistance","elicitation","training"],"pairs-with":[{"skill":"rapport-based-elicitation","reason":"Use it when you need the broader practice surface around rapport-driven extraction and cooperation."},{"skill":"expert-knowledge-elicitation","reason":"Rapport is often the prerequisite for surfacing expert reasoning that resists direct extraction."}],"provenance":{"kind":"legacy-recovered","sourceDocument":"Evaluating the Benefits of a Rapport-Based Approach to Investigative Interviews: A Training Study With Law Enforcement Investigators","sourceAuthors":["Laure Brimbal","Christian A. Meissner","Steven M. Kleinman","Erik L. Phillips","Dominick J. Atkinson","Rachel E. Dianiska","Jesse N. Rothweiler","Simon Oleszkiewicz","Matthew S. Jones"],"sourceArtifact":".claude/skills/evaluating-the-benefits-of-a-rapport-bas/_book_identity.json","importedFrom":"legacy-recovery","owners":["some-claude-skills"]},"authorship":{"authors":["Laure Brimbal","Christian A. Meissner","Steven M. Kleinman","Erik L. Phillips","Dominick J. Atkinson","Rachel E. Dianiska","Jesse N. Rothweiler","Simon Oleszkiewicz","Matthew S. Jones"],"maintainers":["some-claude-skills"]}} |
Evaluating the Benefits of a Rapport-Based Approach
Use this skill when the hidden problem is not lack of instructions but a damaged cooperation pathway. The doctrine here is simple: pressure can force words, but it cannot force truthful, high-yield disclosure.
When to Use
- A human or sub-agent is giving minimal, evasive, or suspiciously polished responses.
- Adding pressure, constraints, or accusation-like framing keeps making output worse.
- A handoff is failing because the receiver does not feel safe, prepared, or respected enough to cooperate.
- The system is over-trusting confidence signals instead of observed behavioral performance.
- You need to reason about why evidence or better tactics are not being adopted in resistant settings.
NOT for Boundaries
This skill is not the primary tool for:
- Straightforward cooperative tasks where the other party is already responsive and truthful.
- Pure technical root-cause analysis when the main problem is system state, not resistance or cooperation.
- Compliance regimes where the design requirement is rigid procedure execution rather than disclosure quality.
- Situations where the operator is looking for intimidation or output-forcing tactics instead of trustworthy information.
Core Mental Models
Coercion Corrupts the Channel
Pressure-based tactics often produce the worst kind of success: output that looks usable but is structurally contaminated by compliance pressure, guesswork, or fabrication.
Rapport Is a Mediated Causal Chain
The chain is tactic -> perceived rapport -> willingness to cooperate -> disclosure. If the cooperation step never forms, "more extraction pressure" is a category mistake.
The Layers Must Stack in Order
Foundation questioning comes first, conversational rapport second, relational rapport third. Skipping the lower layers and jumping to high-trust asks is a design failure, not a shortcut.
Self-Reported Capability Is Not Behavioral Capability
Practitioners and agents often say they understand the tactic while behaving as though they do not. Evaluate the behavior under pressure, not the self-description.
Decision Points
flowchart TD
A[Low-yield or resistant interaction] --> B{Foundation questioning stable?}
B -->|No| C[Fix open questions, listening, and narrative space]
B -->|Yes| D{Perceived rapport improving?}
D -->|No| E[Add autonomy support, empathy, adaptation]
D -->|Yes| F{Cooperation threshold reached?}
F -->|No| G[Use relational rapport and remove pressure signals]
F -->|Yes| H[Request disclosure or higher-stakes detail]
H --> I{Output plausible but suspect?}
I -->|Yes| J[Check for pressure-induced fabrication and reset]
I -->|No| K[Continue with calibrated follow-up]
