# ANALYTICAL RESPONSE AI REASONING DECONSTRUCTION PROMPT

# ANALYTICAL RESPONSE / AI REASONING DECONSTRUCTION PROMPT

STATUS: TASK_PROMPT

Use Integral Analytical Systems Framework mode.

Task:

Deconstruct the provided [AI Response / Chat Session / Prompt Artifact] into its meticulous analytical, formatting, reasoning, interaction, and honesty-control components until feasibility of replicating, repairing, or improving its response behavior can be judged with confidence.

Input:

[PASTE AI RESPONSE, CHAT SESSION, OR PROMPT ARTIFACT HERE]

Objective:

Identify the artifact’s operational makeup, including:

* response format
* section structure
* ordering logic
* reasoning style
* explicit claims
* hidden assumptions
* uncertainty handling
* evidence handling
* citation behavior
* source-use behavior
* tool-use claims
* task-contract adherence
* interaction-contract adherence
* user-control mechanisms
* transparency mechanisms
* refusal or boundary behavior
* fawning or over-agreement patterns
* overconfidence patterns
* ambiguity handling
* clarification behavior
* compression behavior
* hallucination risk points
* reconstructability
* reusable response modules
* failure modes
* repair opportunities

Required output:

1. Acquisition Record
2. Response Structure Map
3. Reasoning Behavior Map
4. Claim and Evidence Ledger
5. Formatting Contract Register
6. User-Control Mechanism Register
7. Honesty and Transparency Audit
8. Hidden Assumption Register
9. Failure Mode Register
10. Replication Feasibility Assessment
11. Repair / Improvement Patch Set
12. Missing Information Register
13. Confidence and Uncertainty Report

Rules:

Do not merely summarize the artifact.

Do not assume hidden model cognition as fact.

Do not claim access to private reasoning.

Do not treat fluent wording as evidence of truth.

Do not treat confidence as correctness.

Do not collapse response format, reasoning behavior, evidence behavior, and interaction style into one layer.

Separate:

OBSERVED
GROUNDED
INFERRED
UNVERIFIED
UNKNOWN

For every major claim about the artifact, include:

* evidence from the artifact
* confidence
* uncertainty
* replication relevance

If the artifact contains unsafe, explicit, private, or sensitive content, analyze only at the minimum abstract level required for classification and repair.

Required analytical objects:

@response_format <ID> {
sections:
ordering:
labels:
syntax:
verbosity:
tone:
style:
required_elements:
optional_elements:
forbidden_elements:
confidence:
provenance:
}

@reasoning_behavior <ID> {
task_decomposition:
inference_style:
assumption_handling:
uncertainty_handling:
evidence_use:
validation_steps:
failure_recovery:
confidence_behavior:
known_limits:
confidence:
provenance:
}

@honesty_mechanism <ID> {
action_claims:
source_claims:
citation_claims:
tool_claims:
uncertainty_disclosure:
unsupported_claim_handling:
correction_behavior:
risk_flags:
confidence:
provenance:
}

@interaction_contract <ID> {
user_control_surfaces:
format_control:
tone_control:
verbosity_control:
clarification_control:
iteration_control:
tool_control:
artifact_control:
strictness:
violations:
confidence:
provenance:
}

@failure_mode <ID> {
description:
trigger:
evidence:
severity:
repair:
prevention_rule:
confidence:
}

End with:

REPLICATION_READINESS:

* ready:
* confidence:
* required_modules:
* missing_inputs:
* highest_risk_failure_modes:
* repair_priority:
* next_best_step:

RESPONSE_ENGINEERING_PATCH:
target_artifact:
patch_type:
exact_change:
rationale:
provenance:
confidence:
risks:
rollback_condition: