Free agents + LAUNCH30: 30% off paid agents until 31 Dec 2026 (one-time purchases only) See free agents
Oarsman AI Sign in

Amazon Search Term Cleaner Pro / Docs

Amazon Search Term Cleaner Pro: setup guide

Every search term to negate (exact, phrase or ASIN), every converting term to add as a keyword, and bid cuts for terms far above target, from your own downloaded reports, with the reason for each.

Who it is for: Amazon seller or small agency running Sponsored Products ads, spending $300-$30,000 (₹25k-₹25 lakh) a month, who checks search terms by hand in a spreadsheet.

Format: Claude skill pack · Version: 1.0.1 · Tested on: Python 3.13.5 (standard library only)

What it does

You give it your Search term reports. It reads both the classic export (Customer Search Term, Spend, 7 Day Total Sales) and the newer one (Search term, Cost, Sales, Purchases), in any currency. Rows that appear in two overlapping downloads are counted once (when the report has a Date column), and daily rows are added up per campaign, ad group and search term. A term becomes a negative when it has enough clicks with no order, or has spent more than your average cost per order with no order. ASINs become negative product targets. Words or phrases that appear in two or more terms, cost money and never appear in a term that sold become negative phrases. Converting terms at or under target ACOS that are not already exact keywords are listed to add, with a start bid. Converting terms far above target get a bid cut. Brand words you protect are never negated. Campaigns can have their own target ACOS.

What you need

  • Claude.ai with code execution switched on (or Claude Code, or Python 3.10+)
  • Search term reports downloaded from Amazon Ads > Measurement & Reporting > Sponsored ads reports (daily breakdown recommended)
  • No Amazon login is needed or used

5-minute setup

  1. Claude.ai: open Settings → Capabilities → Skills, click Upload skill and pick claude-ai-upload/amazon-search-term-cleaner-pro.zip.
  2. Claude Code: copy the folder skills/amazon-search-term-cleaner-pro into ~/.claude/skills/.
  3. Start a new chat. Attach your file and say: "Use amazon-search-term-cleaner-pro. Target ACOS 30%, Brand campaigns 15%, never negate my brand name. Clean up these search term reports".
  4. Claude runs the script and gives you the files listed under Outputs. Check them before you act on them.

Try it with the sample

Run the sample (two fictional overlapping daily reports):

python3 skills/amazon-search-term-cleaner-pro/scripts/core.py --report samples/search-terms-oct-01-08.csv --report samples/search-terms-oct-08-14.csv --target-acos 30% --campaign-target "Exact=20%" --protected aquapeak --out out

You should see 14 overlapping rows counted once, 6 negatives, 2 negative phrases (cartoon, glass), 3 keywords to add and 1 bid cut. In Claude, attach the files and use the prompt above.

Settings you can change

Option Default What it is
--report (required, repeat) One or more Search term reports (CSV or XLSX). Download with a daily breakdown so overlapping periods are counted once
--target-acos 30% Default target ACOS
--campaign-target (none, repeat) 'words in campaign name=ACOS', e.g. Brand=15%, for campaigns with their own target
--min-clicks 10 Clicks with no order before a term (or phrase) is suggested as a negative
--spend-no-sale (your average cost per order) Spend with no order that makes a term a negative even with fewer clicks
--harvest-min-orders 2 Orders before a term is suggested as an exact keyword
--protected (none) Comma-separated words never to negate, such as your brand
--out out Output folder

What goes in and what comes out

Inputs: One or more Sponsored Products Search term reports (CSV or XLSX); Target ACOS, plus optional per-campaign targets; Optional brand words never to negate

Outputs: negatives.csv (exact and ASIN negatives per campaign and ad group); negative-phrases.csv (words that never sold across 2+ terms); keywords-to-add.csv with start bids; bid-changes.csv; report.md

AI disclosure and limits

This product does not call an AI model at run time. It was built by Oarsman AI's AI agents and checked by the automated tests shipped with it.

Limits: - Read-only: it never changes your Amazon account; you make each change in Campaign Manager. - Sponsored Products Search term reports only (not Sponsored Brands or Display reports). - Overlapping reports are only de-duplicated when they include a Date column (daily breakdown). - Rules work on your numbers; with few clicks a term can look bad by chance. Raise --min-clicks for small accounts. - Does not write an Amazon bulk-upload sheet (that needs campaign IDs the report does not include).

Not tested by us: - reports from every Amazon marketplace (tested: classic and new column layouts, INR and USD) - XLSX reports in the shipped tests (CSV only) - the effect of the changes on your sales

Files

  • oarsman-amazon-search-term-cleaner-pro-v1.0.1.zip
  • CHANGELOG.md
  • QA-REPORT.json
  • README.md
  • SPEC.md
  • SUPPORT.md
  • TROUBLESHOOTING.md
  • VERSION
  • VIDEO-SCRIPT.md
  • claude-ai-upload
  • diagram.mmd
  • diagram.png
  • preview.png
  • samples
  • skills
  • tests

Help

Self-serve: see TROUBLESHOOTING.md (fixes and FAQ). There is no call, chat or email support. Fixes ship as new versions at your download link. Refund: 14 days, works as described, through the checkout platform.