CSV Anonymiser

When the user wants to anonymise a CSV file or process new data:

1. Check .gitignore

First, verify that CSV files are excluded from git:

cat .gitignore

If

*.csv

is not present, add it:

echo "*.csv" >> .gitignore

2. Identify the CSV File

The script can auto-detect new CSV files:

• Auto-detection: Files without a timestamp pattern (YYYYMMDD_HHMMSS) in their name
• New files: Any CSV file that hasn't been processed yet
• If multiple new files exist, the script will list them and ask user to specify which one

3. Remove Personal Identifying Information

Columns to remove are configured in

~/.agents/skills/anonymise/config.txt

:

• One column header per line
• Lines starting with

  #

  are comments
• Default includes

  "Record - Airtable ID"

  (PII)

Process the file:

1. Read the CSV file
2. Remove all columns listed in

   ~/.agents/skills/anonymise/config.txt

3. Keep all other columns intact
4. Preserve the header row (minus the removed columns)

4. Add Datetime Stamp

The script will:

1. Create a new file with format:

   YYYYMMDD_HHMMSS.csv

   (just the timestamp)
2. Delete the original file

Example:

ILR data 25_26-Social value data for Dingley.csv

becomes

20231216_143022.csv

5. Run the Anonymisation Script

Execute the Python script located in the same directory as this skill:

Auto-detect mode (finds CSV files without timestamps):

python3 ~/.agents/skills/anonymise/anonymise.py

Specific file mode:

python3 ~/.agents/skills/anonymise/anonymise.py "filename.csv"

If multiple new files are found in auto-detect mode, the script will list them and you'll need to specify which one to process.

6. Confirm Completion

After processing:

1. Confirm the anonymised file was created with timestamp
2. Confirm the original file was removed
3. Confirm .gitignore includes

   *.csv

4. Remind user that CSV files won't be committed to git

Notes

• Uses British English spelling throughout
• Handles UTF-8 BOM encoding (

  utf-8-sig

  ) for compatibility
• Preserves all data except the first column
• Assumes consistent CSV structure across files
• Files are automatically excluded from version control