GITHUB
AUDITFIXMEASURE DAY 01 OF 30 — LIVE
V0.1 // FREE // OPEN SOURCE // MIT // 0 PIP INSTALLS

I gave Claude the keys to my Amazon store.

This isn't a demo. Claude Code has been running this store's ops for a while now — audits, listing edits, review requests, tracking — through the same APIs Amazon gives every seller. $130K sold, a quarter-million dollars of stock under management. 455 review requests on autopilot, 4 broken listings found and patched. Today I'm open-sourcing the whole setup, and publishing the next 30 days of numbers here, win or lose.

$ git clone https://github.com/EbrahimElbagory/claude-fba.git
MACOS · LINUX · WINDOWS // PYTHON 3.9+ STDLIB // GET API KEYS →
$130K
SOLD — TRAILING 12 MONTHS
$250K
STOCK UNDER MANAGEMENT — AT LIST
455
REVIEW REQUESTS — ON AUTOPILOT
80K
SESSIONS ANALYZED
28
PRODUCTS CONNECTED
$0
FOREVER. MIT LICENSED.
// 01 — THE BET

The public scoreboard.

Every change ships against a baseline. Current cycle's was captured 2026-07-22; the Day-30 column fills in public as weekly numbers land. No projections, no cherry-picking.

FOLLOW THE NUMBERS ON X →
METRIC
DAY_0
DAY_30
STATUS
Reviews — newest ASIN
13
?
MEASURING
Reviews — 2nd-newest ASIN
51
?
MEASURING
Conversion — the fixed listing
2.7%
?
MEASURING
Core keyword rank
#6–9
?
MEASURING
Backend keywords filled
0/4
4/4
DONE — DAY 0
LIVE EXPERIMENT ON A REAL STORE // WEEKLY SNAPSHOTS MON 09:00
Want this running your store?

Drop your email — first in line when this becomes a product you can point at your own store. Amazon today; Shopify and eBay decided by demand. No newsletter, no spam.

// 02 — THE PATIENT

The store, anonymized.

Which store? Doesn't matter — the numbers do. Every figure below is pulled straight from the seller APIs. Identity withheld so the experiment stays clean; specs published so you can judge whether this scale means anything to yours.

SPEC
VALUE
NOTE
Gross sales
$130,195
TRAILING 12 MONTHS
Units sold
6,454
AOV ~$21 · 7.7% CVR
Take-home before COGS
47.9%
~19% NET MARGIN
Inventory on hand
$250K at list
7,005 UNITS · $132K AT TODAY'S PRICES · $95K AT COST
Traffic analyzed
80,803 sessions
115K PAGE VIEWS
Products connected
28
19 DEEP-AUDITED EVERY CYCLE
Ads
Coming soon
AD-MANAGER SKILL — IN LIVE TESTING
REAL STORE. REAL MONEY. IDENTITY WITHHELD — NUMBERS AREN'T.
// 03 — THE EVIDENCE

The ops log.

One real session from the log — 2026-07-22, my own store. This is what a routine audit-and-fix cycle looks like when the operator is an agent with API access. Every number came straight out of an API response, nothing is a projection. Caught that session: an estimated $3.5–9K/yr in process leaks.

CLAUDE-FBA — SESSION 2026-07-22 — VERBATIM
14:07audit: 19 ASINs · ~25 API calls · listing quality, keyword ranks, BSR, reviews, inventory
14:31WARN backend search terms empty on 4/4 audited listings
14:32WARN one listing's bullets describe a different product · converting 2.7% vs 10–14% siblings
15:58listings PATCH ×4 → ACCEPTED (titles + backend keywords · backups saved first)
16:44solicit_reviews.py → 437 sent, 18 ineligible, 0 errors
16:59rerun (dedupe test) → 0 sent, 452 already logged
17:03cron installed: reviews daily 10:00 · impact snapshot mon 09:00
17:10baseline captured → est. $3.5–9K/yr in process leaks documented
nowmeasuring
// 04 — THE WIRING

Two data planes. One brain. A cron.

The market layer runs on Helium 10's official MCP (paid plan). The store layer runs on a free self-authorized SP-API app you create once, in ~30 minutes. SETUP GUIDE →

[MARKET LAYER]
HELIUM 10 MCP

Keyword volume, ranks, listing quality scores, CPR, competitor intel.

[STORE LAYER]
YOUR OWN SP-API

Listings, orders, review solicitations, reports. A ~300-line stdlib client — no pip installs.

[PERSISTENCE]
LAUNCHD / CRON

Daily review requests 10:00, weekly impact snapshots Mon 09:00.

// 05 — THE CONTENTS

Five skills. One loop. A sixth on the bench.

SKU_01[AUDIT]
Store-Audit

A full checkup of every listing you own: content quality, which keywords you rank for and where, sales-rank trend, reviews, and months of stock left. Each audit is compared to the last one, so you see exactly what moved.

READ THE SKILL →
SKU_02[REVIEWS]
Review-Automation

Amazon lets you request one review per order, day 5–30. This sends that official request for every order you've missed (my catch-up: 437), then runs every morning so no order is missed again. Can't double-send.

READ THE SKILL →
SKU_03[EDITS]
Listing-Guard

Edits live listing text (titles, backend search terms) safely: checks what Amazon allows on your product type, drafts the copy, changes nothing until you approve the exact wording. Old version saved first — undo is one command.

READ THE SKILL →
SKU_04[MEASURE]
Impact-Tracker

Proves whether changes worked. Records every number before you touch anything, then logs traffic, conversion, units, and revenue per product weekly. You compare against a baseline instead of guessing.

READ THE SKILL →
SKU_05[KEYWORDS]
Keyword-Research

Finds keywords worth going after: what competitors rank for that you don't, expanded seed terms, all sorted by how cheap each is to win. Ends with a map — these words go in your title, bullets, backend, and ads.

READ THE SKILL →
SKU_06[COMING SOON]
Ad-Manager

The ads loop: pulls every campaign, finds search terms burning money with zero sales, flags winners stuck in broad match, and drafts bid, budget, and negative-keyword changes — nothing spends a cent until you approve it. In testing on my store now.

SHIPS AFTER LIVE TESTING →
// 06 — HANDLING INSTRUCTIONS

The safety model is the feature.

Letting an agent touch a revenue-producing store only works with hard rules. This store is live cargo — handled accordingly.

[✓]
Read-only by default

A full audit writes nothing. Most sessions never mutate anything.

[✓]
Draft → human approval → execute

No buyer-visible change ships until you approve the exact copy in chat.

[✓]
Backup before write, verify after

Pre-edit state saved per SKU; post-PATCH re-read checks for new errors. Rollback is one command.

[✓]
Official APIs only

Your own SP-API app, Helium 10's own MCP. Nothing scraped, no ToS games.

[✓]
Honest about limits

No API path for something? The skills say so and hand you a manual checklist — instead of pretending.

// 07 — DECLARATIONS

Questions sellers actually ask.

Is this against Amazon's terms of service?

No. It's your own store accessed through your own self-authorized SP-API app — the same official interface every seller tool uses. Review requests go through Amazon's Solicitations API, functionally identical to the "Request a Review" button in Seller Central. Nothing is scraped.

Do I need Helium 10?

Only for the keyword side — search volumes, rank tracking, competitor keyword data (their official MCP, paid plan). Everything else needs nothing but the free SP-API app you make yourself: review automation, listing edits, impact tracking, sales analytics, the whole loop.

What does it cost to run?

Claude Code (your existing subscription) + optionally Helium 10. The toolkit itself is free, MIT-licensed, stdlib-only Python — no servers, no pip installs. A full audit uses ~25 of your 1,000 monthly Helium 10 MCP calls.

Can it break my listings?

The harness is built so it can't do so silently: schema checked before edits are promised, every change drafted for your approval, pre-edit state backed up before any write, result verified after. Rollback = re-apply the backup.

Windows / Linux?

Scripts and skills run anywhere Python 3.9+ does. The two automation templates are macOS launchd; use cron on Linux or Task Scheduler on Windows.

What's on the roadmap?

Amazon Ads API campaign automation (CPR-sized ranking pushes — the auth client already ships in scripts/), Search Query Performance funnel analysis, multi-marketplace support. Issues and PRs welcome.