Daily new listings watch on competitor Amazon storefronts

Build me an agent workflow that watches a curated list of competitor Amazon seller storefronts and flags brand-new product listings every morning.

Trigger: cron, every weekday at 7:00 AM America/New_York.

Inputs I want to configure once (workflow config):

- Google Sheet spreadsheet ID (used for all reads and writes).

- Slack channel ID or name to post the daily summary to.

- Amazon marketplace country code (default US).

The Google Sheet has three tabs the agent needs to know about:

1. Sellers: columns A = seller_id, B = optional label like "brand" or "wholesale". Header row on row 1, data starts row 2.

2. Seen ASINs: state tab. Columns A = seller_id, B = ASIN, C = first_seen_date. This is our memory of every product we have ever pulled per seller.

3. New Listings: append-only log. Columns A = date, B = seller_name, C = seller_id, D = ASIN, E = title, F = price, G = category, H = star_rating, I = best_sellers_rank, J = amazon_url.

Step-by-step behavior on each run:

1. Read the Sellers tab with Google Sheets Get Values. Read the Seen ASINs tab with Google Sheets Get Values and build an in-memory set of (seller_id, ASIN) pairs and a set of seller_ids we have ever pulled before.

2. For each seller_id in the Sellers tab, call Real-Time Amazon Data List Seller Products and paginate through the seller's full current storefront. Collect every ASIN the seller is currently listing.

3. Diff: for each seller, keep only ASINs that are not already in the Seen ASINs set for that seller. These are the genuinely new listings for today. If the seller has zero new ASINs, skip the seller entirely (no enrichment, no writes, no Slack line).

4. For each new ASIN, call Real-Time Amazon Data Get Product Details to enrich with title, price, primary category, star rating, and Best Sellers Rank.

5. If this is the first time we have ever seen this seller (they have zero rows in Seen ASINs), also call Real-Time Amazon Data Get Seller Profile once so we have a readable seller_name to use in the sheet and Slack. Otherwise, reuse the seller_name from the most recent New Listings row for that seller_id, and if none exists, fall back to the seller_id.

6. Append one row per new listing to the New Listings tab with Google Sheets Append Values, using today's date in America/New_York.

7. Append one row per new (seller_id, ASIN) pair to the Seen ASINs tab with Google Sheets Append Values so tomorrow's diff is accurate. Never overwrite or delete existing Seen ASINs rows.

8. Post a Slack message with Slack Bot Send a Message to the configured channel. The message should include: total new listings today across all sellers, a per-seller line count (only for sellers that had at least one new ASIN), and a bullet list of the top three most notable launches. "Most notable" means highest star rating, best (lowest number) Best Sellers Rank, or an obvious category expansion for that seller. Include a one-line reason for each highlight and a link to the Amazon product page. Use Slack mrkdwn (single asterisks for bold, angle-bracket links).

Error handling: if List Seller Products fails for a specific seller_id after a couple of retries, skip that seller, do not write anything for them, and add a short "errors" section at the bottom of the Slack message listing which seller_ids failed and why. A single bad seller should never abort the run.

Idempotency: state lives entirely in the Seen ASINs tab. If the run is re-triggered on the same day, sellers with no new ASINs should be silently no-ops.

If there are zero new listings across every seller today, still post a short Slack message like "No new listings today across N sellers watched" so the team knows the check ran.