Flowhub MCP skills make it quick and easy to complete detailed, multi-step workflows in your store with an AI agent. In this article, we'll guide you through using skills so your team can work more consistently across products, inventory, pricing, deals, and analytics.
What is a skill?
A skill is a predefined, reusable set of instructions and supporting context that teaches an MCP client how to complete a specific Flowhub task. Skills add structure about what to do, what context to gather, and what order to follow, which helps with workflows that are detailed or repetitive. Because skills are reusable, teams can package their own best practices into something an agent can apply again and again.
Before you begin, connect the Flowhub MCP server and authenticate with your Flowhub account. Visit the Flowhub MCP Setup Guide to learn more.
First time adding a skill? Here's a quick start guide to adding skills.
First time adding a skill? Here's a quick start guide to adding skills.
Get the Flowhub skills
There are two ways to get the Flowhub skills. Select whichever source your team prefers. Google Drive is a quick download, and Git is useful for version control and team sharing.
View the skills in our MCP skill repository in GitHub.
Upload the skill to your AI
Add the skill to your AI client using the method your client supports. Depending on the client, you might add the skill as a file, paste its instructions, or point the client at the Git link. Check your client's setup instructions to confirm how it expects skills to be added.
Confirm the skill is available in your client.
In clients that list installed skills, type / to see them, or ask your agent, "What Flowhub skills do I have available?"
Use a skill in a prompt
Describe the outcome you want in plain language and let the client apply a skill you have added. If your client supports skills, it may automatically apply a relevant skill when your prompt matches that skill's description.
For example, ask your client to add a new product, adjust pricing, build a deal, or summarize last month's sales.
The client decides whether a skill is relevant based on your request and the skill's description.
đĄ Tip: This approach keeps the interaction conversational. Focus on the outcome rather than the exact skill name.
Invoke a skill directly
You can call a specific skill when you already know which workflow you want the client to follow. Type / followed by the skill name to call that skill.
Direct invocation is helpful for confirming that a skill you added is available and working as expected.
Check your client's setup instructions, because invocation varies by client. Some clients treat skills as slash commands, some surface them in a menu, and others apply them automatically based on the conversation.
Flowhub Skills Library
Deal Optimizer
Scans your last 30 days of product performance and recommends which products to put on a deal, then measures whether your existing deals are paying off.
How it works
Ask your agent what to discount or where you're over-discounting. Example: "What should I put on sale to move slow inventory this month?"
Ranks slow-moving products that have margin room and aren't already discounted, and also flags winners to leave alone and deals that aren't earning their keep.
Backs its picks with evidence by comparing gross profit per day on deal-on versus deal-off days, rather than assuming a discount worked.
Read-and-recommend: it suggests deals but creates them one at a time only after you confirm, and never below cost or a brand or state price floor.
Uses the analytics_sales, analytics_inventory, manage_deals, and manage_price_profiles tools.
How to get it
View on GitHub