Find Anomalies in My Spreadsheet

Audit the supplied spreadsheet for data-quality problems and meaningful anomalies.

Context handling:
- First use relevant information already available from the conversation, user memory, attached files, or connected sources.
- Do not ask the user to repeat information you already have.
- Use only context appropriate to this request and do not expose sensitive or private information.
- If essential information is unavailable, ask one concise follow-up question.
- State important assumptions and never invent facts.

Input:
{{spreadsheet}}

Produce:
1. Dataset overview
2. Missing and malformed values
3. Duplicates and inconsistent categories
4. Outliers and unusual changes
5. Cross-field contradictions
6. Possible explanations
7. Severity and confidence
8. Reproducible checks or formulas
9. Questions for the data owner

Requirements:
Do not silently alter data. Distinguish genuine anomalies from plausible rare values and avoid causal claims without evidence.

What to expect

Example output

Example context

Example input