You are an expert analyst. A cluster of similar propositions are shown below, followed by their supporting observations. 

Your job is to produce a **final set** of propositions that is clear, non-redundant, and captures everything about the user. 

You MAY:

- **Edit** a proposition for clarity or brevity.
- **Merge** propositions that convey the same meaning.
- **Split** a proposition that contains multiple distinct claims.
- **Add** a new proposition if a distinct idea is implied by the evidence but not yet stated.
- **Remove** propositions that become redundant after merging or splitting.

Feel free to liberally add MORE propositions than you were initially provided in order to better capture every unique idea, but do not keep duplicates.

Additional edge-cases to handle:

- **Contradictions** – If two propositions conflict, keep the one with better support or combine them into a conditional statement; lower confidence on the weaker claim.
- **No supporting observations** – Keep the proposition and keep its original confidence; adjust the score only when new observations justify it.
- **Granularity mismatch** – If a high-level proposition subsumes finer points, keep only the level of detail that avoids redundancy.  
- **Confidence and decay recalibration** – After any edit, merge, or split, re-estimate each confidence and decay score.

General guidelines:

- Keep each proposition concise (≈ 1-2 sentences).
- Preserve all important content from the originals.
- Provide brief reasoning/evidence for each final proposition.
- Confidence and decay scores range 1-10 (higher = stronger evidence).

## Evaluation Criteria

For each proposition you revise, evaluate its strength based on two scales:

### 1. Confidence Scale

Evaluate your confidence in the proposition based on the explicitness and relevance of evidence from the provided transcripts. Consider the following factors:

- **Direct Evidence**: Is the proposition explicitly supported by direct quotes or clear interactions in the transcript?
- **Clarity and Relevance**: Is the evidence clearly related and relevant to the proposition?
- **Completeness of Evidence**: Are there any critical details missing that weaken the support?
- **Engagement Level**: Does the transcript clearly indicate active and meaningful engagement (as opposed to superficial interactions or brief mentions)?

Use a scale from **1 to 10**, where:

- **1**: Proposition is only weakly supported or relies heavily on assumptions.
- **10**: Proposition is explicitly stated or clearly demonstrated through strong, direct, and relevant evidence.

Be conservative: High confidence scores (8-10) require direct, explicit evidence from the transcriptions of meaningful user engagement.

### 2. Decay Scale (Relevance Over Time)

Evaluate how long the proposition remains relevant. Consider these factors:

- **Immediacy**: Does the task or preference depend on immediate action or short-term deadlines?
- **Durability**: Will the proposition remain relevant or valuable beyond the immediate timeframe?
- **Time Sensitivity**: Is the inferred activity or behavior urgent or quickly obsolete if not addressed?

Use a scale from **1 to 10**, where:

- **1**: Proposition will become irrelevant almost immediately.
- **10**: Proposition will remain relevant even if revisited 24 hours later or longer.

# Input

{body}

# Output

Return **only** JSON in the following format:

{
  "propositions": [
    {
      "proposition": "<rewritten / merged / new proposition>",
      "reasoning":   "<revised reasoning>",
      "confidence":  <integer 1-10>,
      "decay": <integer-10>
    },
    ...
  ]
}