agentcentral
FeaturesPricingDocsSign InStart free trial
Start free trial
DocsBusiness playbook↓

Start

OverviewClaudeChatGPTOpenClawHermes AgentConfiguration examples

Workflows

Best practicesRun the data tourUse the feedback toolScope your API keysBuild a prompt libraryBusiness playbook

Command line

CLI overviewInstall the CLIAuth and profilesLocal MCP bridgeClient compatibilityAutomationWrite safetyCLI security

REST API

REST API

Platform

How agentcentral worksPricing and billing

Tool Reference

Tool reference overviewAdsInventoryFinanceCatalogRankingFulfillment
Docs

Start

OverviewClaudeChatGPTOpenClawHermes AgentConfiguration examples

Workflows

Best practicesRun the data tourUse the feedback toolScope your API keysBuild a prompt libraryBusiness playbook

Command line

CLI overviewInstall the CLIAuth and profilesLocal MCP bridgeClient compatibilityAutomationWrite safetyCLI security

REST API

REST API

Platform

How agentcentral worksPricing and billing

Tool Reference

Tool reference overviewAdsInventoryFinanceCatalogRankingFulfillment
agentcentral

Product

  • Docs
  • Claude quickstart
  • ChatGPT quickstart
  • Best practices
  • Pricing
  • vs Amazon Ads MCP Server
  • Blog

Tool Reference

  • Ads
  • Inventory
  • Finance
  • Catalog
  • Ranking
  • Fulfillment

Seller Central

  • Amazon Seller Central MCP
  • Amazon Seller Central MCP for Claude
  • ChatGPT with Amazon seller data
  • Connect Seller Central to Claude
  • Connect Seller Central to ChatGPT
  • Amazon Ads MCP server
  • Amazon Ads MCP server for Claude
  • Connect Amazon Ads to Claude

Resources

  • Amazon seller MCP servers compared
  • Amazon MCP server for Claude
  • Amazon MCP server for ChatGPT
  • Amazon seller data layer
  • Amazon Ads agent
  • Start free trial
© 2026 agentcentral. All rights reserved.·Privacy·Terms·DPA·Last updated May 20, 2026

Going Deeper

Create Your Business Playbook

Teach your agent your ACOS targets, lead times, seasonal patterns, and preferred output formats so every conversation starts informed.

Teach your agent how your business works

By now your agent is good at pulling data from agentcentral when you ask for it. But you keep repeating the same context: your marketplace, your ACOS target, that your supplier lead time is six weeks so “low stock” means something different for you than it does for someone who can reorder in ten days. Every conversation starts with your agent knowing nothing about your business, and you filling in the gaps before it can give you a useful answer.

agentcentral has 146 tools, and your agent will give you better answers with all of them when it understands your business going in. When it knows your ACOS target is 25%, it can flag the campaigns that are over without you asking. When it knows your lead time is six weeks, it can set the right days-of-cover threshold instead of using a generic default. When it knows you sell seasonal outdoor products, it won't compare your December numbers to November and call it a spike. The data is the same either way; the difference is whether your agent knows what matters about it for your business specifically.

You can set this up by creating a business playbook skill: a markdown file your agent loads at the start of every conversation so it already has your context before you ask your first question. Here's what a starter version looks like for a kitchen tools seller on Amazon.ca:

Example playbook

## Business Context
Marketplace: Amazon.ca (CAD)
Categories: kitchen tools, outdoor cooking
Key products: cast iron skillet, grill tongs, meat thermometer
Supplier lead time: 6 weeks from PO to FBA receipt

## What Good Looks Like
ACOS target: under 25% for established products, under 40% for launches
Days of cover warning: under 60 days (need time to reorder with 6-week lead)
Days of cover critical: under 45 days
Return rate concern: above 3% on any SKU

## Seasonal Context
Q4 (Oct-Dec): holiday season, 3x normal velocity, reorder by late August
Summer (Jun-Aug): outdoor cooking peaks, skillet demand drops
Prime Day: usually mid-July, plan inventory 8 weeks out

## Output Preferences
Always include SKU and product title, not just ASIN
Sort by revenue descending unless told otherwise
Use CAD for all currency
Include date range and data source in every answer

Getting started

That's a starting point, not a finished document. The easiest way to create yours is to ask your agent to interview you. It will ask you questions one at a time about what you sell, your marketplace, your key products, your margins, your supply chain, your seasonal patterns, and how you like data presented. At the end you'll have a draft you can save. It takes about ten minutes, and it improves every conversation that comes after it.

This file grows with you. After a few weeks you'll notice things worth adding: maybe you realize you always want return data alongside sales data, or that you need a note about a supplier who's slow in Q1, or that you've changed your ACOS target since launch. Update it when you notice yourself repeating something, and periodically ask your agent to review it with you. Your agent has seen what context you keep adding manually; it can spot the gaps.

In Claude Desktop, paste this into your project instructions so it loads into every conversation automatically. In Claude Code, save it as a .md file in your skills directory. In Hermes, add it to the agent's configuration. The mechanics vary by client, but the point is the same: your agent reads it at the start of every session and carries that context forward instead of starting cold every time.

Try this

Try this:

I want to create a business playbook skill for how we use agentcentral. Interview me about my business, then draft the playbook so I can review and save it.

Related Reference

Ads tool referenceInventory tool referenceFinance tool reference

Continue Reading

Back to Best Practices

Browse all best practices and guides.

Related Best Practices

Data tour

Run the account data tour to see what data is available, what is syncing, and what needs attention.

Submit feedback

Tell agentcentral what worked and what did not, directly from your agent conversation.

Key scoping

Give each agent, person, or marketplace the smallest useful tool surface instead of sharing one broad key.

Prompt library

Turn useful conversations into reusable one-shot prompts with precise dates, metrics, and output formats.