- Home
- Skills
- Data & Databases
- Check a Spreadsheet Migration Lost Nothing
More screenshots
Works with the AI Agents you already use
Check a Spreadsheet Migration Lost Nothing
Before you sign off on moved data, this checks the new spreadsheet against the old one, record by record — and flags doubts instead of guessing.
$9
Check a Spreadsheet Migration Lost Nothing
Example session with this skill installed
Audit these two CSVs. Use 'Legacy_ID' as the ID column and check the 'Description' field for line breaks. Use this field map: {"Company": "Account Name", "Email": "Primary Contact"}.
- Read your context and instructions
- Compiled the check a spreadsheet
- Generated the data export
I have completed the migration audit between source_export.csv and migrated_data.csv. The audit failed due to 3 missing records and 2 field mismatches.
Migration Audit Report
| Check Type | Result | Findings |
|---|---|---|
| Record Counts | FAIL | Source: 1,050 / Destination: 1,047 |
| Duplicate IDs | PASS | 0 duplicates found |
| Missing Records | FAIL | 3 IDs: CUST-882, CUST-901, CUST-112 |
| Unexpected Records | PASS | 0 records |
| Field Mismatches | FAIL | 2 mismatches (see below) |
| Multiline Notes | PASS | 0 line breaks found in 'Description' |
Field Mismatch Details:
ID: CUST-445: Field 'Company' expected 'Acme Corp', found 'Acme Corporation'ID: CUST-512: Field 'Email' expected 'j.doe@web.com', found ''
Potentially Lost Values:
ID: CUST-602: Source value 'Gold Tier' not found in destination record.
Next steps
- Investigate the 3 missing records in the migration script log.
- Update the field map or migration logic to account for 'Acme Corporation' naming convention.
- Re-run the audit once the destination file has been regenerated to achieve a PASS status.
source_export.csv
CSV · data export
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Migrations fail quietly. A record vanishes, two values swap columns, a long note gets cut off — and nobody notices until the old system is gone. This skill is the audit you run before delivery, comparing the source export against the migrated file, record by record. Its core discipline is simple: the audit flags, humans decide. It never guesses where ambiguous content belongs, and it never fixes data itself.
Inside the package: the skill file with the six checks and the field-map method, a pre-delivery checklist, an example audit report in JSON — the report format is the deliverable you keep with the migration as proof of what was checked — and the checker program itself, a single Python file in the reference folder that runs the six checks and writes the report. The six checks: record counts on both sides; duplicate record IDs in either file; records missing from the new file; records in the new file that shouldn't exist; values that can't be found anywhere in the matching new record; and notes cells that still contain line breaks where the target format needs them flat. A field map — a small file pairing each old column name with its new one — adds the check that presence alone misses: whether each value landed in the right field, compared exactly, character for character.
A run goes like this: freeze both files (no edits between audit and delivery), agree the ID column and the field map in writing, and the agent runs the audit. Anything ambiguous lands in a review queue with the record ID and the evidence, and a person marks each item a real loss, an accepted change, or a false alarm. Expected reformatting — dates rewritten, phone numbers respaced — will show up as mismatches; that's the design, not a malfunction. Fixes happen in the migration, never by hand-editing the target file, and the full audit reruns on the corrected export. Delivery happens only on a clean pass, with the saved report attached. Scope: this package is the audit method, the checklist, the report format — and the checker itself: a complete Python program (data_mapping_qa.py, nothing else to install) is bundled in the reference folder, so the comparison runs as a program, not an eyeball.
Requires UTF-8 CSV inputs with unique headers. Merged or split fields are out of scope. Human review is required to clear the audit queue; the tool does not "learn" to ignore mismatches.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 6 days ago