CSV Field Inspector

The Field Inspector is Workhog's secret weapon for spreadsheet data. Instead of describing your data to AI, just click a cell — the inspector automatically tells Claude the type, pattern, sample values, and full row context.

Why It's 10x Better Than Prompting

Without Inspector

You:

"I have a CSV with phone numbers in column C. They're in different formats. Can you standardize them?"

Claude:

"What formats are they in? Can you share a few examples? How many rows? What column name?"

You:

"Like 603-555-0174, (207) 555-0198, 5551234567... about 85 rows. Column is called Contact."

Claude:

"Got it. Which format do you want? (XXX) XXX-XXXX or XXX-XXX-XXXX?"

4 messages just to explain context

With Inspector

You click cell C2. Inspector shows:

Cell:C2
Column:Contact
Value:603-555-0174
Type:
text
Pattern:
phone
Samples: 603-555-0174, (207) 555-0198, 5551234567, ...
85 rows in column

You:

"Standardize these phone numbers"

Claude:

"Done. All 85 phone numbers in Contact column formatted as (XXX) XXX-XXXX."

1 click + 1 message. Done.

The golden rule

Click, don't describe. Every time you click a cell, Claude instantly knows the value, type, column pattern, 10 sample values, and the full row context. That's more information than you could type in 5 messages.

Example

Try it yourself:

Upload any CSV or Excel file, open it in the viewer, and click a cell. The Field Inspector panel will appear on the right showing all the data Claude can see.

Cell Selection

When you click any cell in the CSV viewer, the Field Inspector panel opens on the right side. It displays:

What Claude Sees When You Click a Cell

Cell reference — A1 notation (e.g., B5)
Cell value — the raw content
Cell type — text, number, date, boolean, or empty
Column name — the header of this column
Column pattern — auto-detected semantic type
Sample values — 10 representative values from the column
Row context — all values in the same row
Formula detection — shows raw formula if cell has one

All of this information is sent to Claude automatically when you type your next message. You don't need to explain anything — Claude already has full context. Compare this to manual prompting in the Best Practices guide.

Column Pattern Detection

The inspector doesn't just see "text" — it recognizes semantic patterns. When it detects that a column contains phone numbers, currency values, or email addresses, it reports the pattern so Claude can make smarter decisions.

currency

$1,234.56

Dollar amounts with symbols

phone

(603) 555-0174

US phone numbers, any format

email

jane@acme.com

Email addresses

date

2026-01-15

Dates in various formats

percentage

42.5%

Percentage values

url

https://...

Web URLs

ssn

***-**-1234

Social security numbers

zip

90210

ZIP/postal codes

numeric

1234567

Plain numbers

Pattern detection happens automatically when you click a cell. The pattern is inferred from the first 10 non-empty values in the column. This means Claude knows "this column contains phone numbers" without you saying a word.

Quick Actions

The inspector offers scope-aware quick actions — pre-built transformations matched to the type of data you've selected. These are one-click operations that don't require any prompting.

Cell-Level Actions

  • trim_whitespace — remove leading/trailing spaces
  • uppercase / lowercase — change case

Column-Level Actions

  • format_currency — normalize to $X,XXX.XX
  • format_date — normalize to YYYY-MM-DD
  • format_phone — normalize to (XXX) XXX-XXXX
  • fill_empty — replace blanks with N/A or a default
  • sort_column_asc / sort_column_desc — sort the data
  • deduplicate — remove duplicate values

Tip

Quick actions know their scope. When you click a single cell, you get cell-level actions. When you select a column header, you get column-level actions that apply to every row. No ambiguity about "which cells?"

Smart Suggestions

When you collect multiple cells, the inspector offers intelligent suggestionsbased on your selection pattern:

Cross-sheet matching

If you select cell B5 from Sheet1, the inspector suggests matching B5 from Sheet2 and Sheet3. Useful for comparing values across periods or entities.

Error cell collection

If you select a cell with a #DIV/0! error, the inspector finds all other error cells and offers to add them to your selection. Fix all errors in one operation.

Same-column matching

Collecting cells from the "Amount" column? The inspector suggests adding the Amount column from other sheets too.

Google Sheets Mode

When you export a CSV to Google Sheets, the inspector gains live inspectioncapabilities. It monitors what you copy from the Google Sheets iframe and automatically matches it back to the source data.

How It Works

  1. Upload a CSV to your workspace
  2. Click "Edit in Google Sheets" in the CSV viewer toolbar
  3. The file opens in an embedded Google Sheets iframe
  4. Copy any cell or range from Google Sheets
  5. The inspector automatically detects what you copied and shows the Field Inspector panel

Three-Pass Matching

Matching clipboard data back to source cells is tricky — Google Sheets may display numbers differently than the raw CSV. The inspector uses a three-pass matching algorithm:

1

Exact match

Clipboard text matches cell value exactly

2

Formula-aware

Allows formula cells where the display value differs from stored value

3

Numeric tolerance

Handles rounding (e.g., CSV has 24.42, Sheets displays 24.00)

The inspector also detects Google Sheets formula errors: #DIV/0!, #N/A,#NAME?, #NULL!, #NUM!, #REF!,#VALUE!, #SPILL!, #CALC!, #ERROR!. Each error gets a human-readable description so Claude can diagnose the issue.

One-Click Actions (Learned from You)

Workhog learns from your behavior. When you've formatted phone numbers before, the next time you select a similar cell, the inspector shows one-click actions based on your history:

One-Click Actions

Format as Currency

89% similar to cells you've fixed · Used 24 times

1-click

Trim Whitespace

76% similar · Used 12 times

1-click

The match score tells you how relevant each suggestion is. Actions with higher scores mean the current cell is very similar to cells you've operated on before. The system uses Workhog's deep repo learning to match based on:

  • Cell type and column pattern (phone, currency, date, etc.)
  • Column name and position
  • What you did to similar cells in the past
  • Session context (recent operations)

Multi-Sheet Support

Excel files with multiple sheets are fully supported. The inspector tracks which sheet each cell comes from, and smart suggestions offer cross-sheet operations like:

  • Comparing values across sheets (e.g., "Q1 vs Q2 revenue")
  • Collecting the same cell from all sheets
  • Detecting cross-sheet formula errors

Sheet names appear as tabs in the CSV viewer. Click a tab to switch sheets, then click any cell to inspect it. The inspector always shows which sheet the current cell belongs to.

Best Practices

1

Always click before asking

Click a representative cell before typing your request. Claude will have all the context it needs.

2

Use column headers for bulk operations

Click a column header to select the entire column. Quick actions like format_currency apply to all rows.

3

Collect cells for batch operations

Select multiple cells to build a collection. Smart suggestions help you find related cells across sheets.

4

Let patterns guide your prompts

If the inspector shows "pattern: currency" for a column, Claude already knows the format — just say "standardize" instead of explaining the format.

5

Use Google Sheets for complex edits

Export to Google Sheets for formula-heavy work, then use clipboard inspection to tell Claude about specific cells. See Integrations for Google Drive setup.

What's Next

CSV Field Inspector - Workhog Help