- Home
- Skills
- E-commerce
- Return & Refund Reply Guard
Works with the AI Agents you already use
Return & Refund Reply Guard
Check return policies, dates, refund amounts, and approval evidence. Get a clear internal review and an unsent customer reply.
$4
Return & Refund Reply Guard
Example session with this skill installed
Review this return case and draft an English reply. Do not send anything.
The applicable policy says accepted returned items receive their recorded net price plus allocated tax. Original outbound shipping is not refundable. The returned mug has been accepted. Its net paid price after the recorded item discount is 24.00 USD, and allocated tax is 1.92 USD. I am the store owner and have approved 25.92 USD total for this mug. The current refund ledger shows 10.00 USD completed and a separate 15.92 USD pending on the same mug. No bank-arrival estimate is available.
Customer: “I only got ten dollars. Please refund the rest today.”
- Read your context and instructions
- Compiled the return & refund
Internal review
- Policy fit: Supported under the supplied policy for the accepted mug. The eligible components are its recorded net price and allocated tax: 24.00 + 1.92 = 25.92 USD. Original outbound shipping is excluded by the supplied policy.
- Approval: Documented in the store owner's instruction for 25.92 USD total, not an additional 25.92 USD.
- Execution: The current ledger records 10.00 USD completed and 15.92 USD pending for this mug. Together they equal the approved total. The customer reports receiving the completed portion; receipt of the pending balance is not established.
- Next step: Verify the pending transaction's outcome before considering any new refund. A second payment now could duplicate the balance. There is no supported bank-arrival date to promise.
Customer reply (unsent draft)
The approved refund for your returned mug is $25.92 USD, including its allocated tax. Our refund record shows $10.00 completed and the remaining $15.92 still pending. That remaining amount is not yet confirmed as complete, so I can't confirm that it will arrive today or give a reliable arrival date from the information available.
Send check: Ready for merchant review. This is an unsent draft; no additional refund has been submitted.
Connects securely to your tools. The creator never sees your data.
About this skill
A customer wants a refund. Your policy mentions a return window, the order has a discount, and someone says the refund was already approved. What can you safely say back?
Return & Refund Reply Guard turns the policy, order evidence, and customer message you provide into a concise internal review and a practical, unsent reply. It keeps three questions separate: does the request fit the supplied policy, has the remedy actually been approved, and what action is evidenced in the records?
Use it for:
- Return-window checks, including original request dates and time-zone boundaries.
- Missing return-label details, postage responsibility, and inspection conditions.
- Partial refunds involving quantities, recorded discounts, tax, shipping, or earlier refunds.
- Refund-status replies that distinguish a pending transaction from completed payment.
- Damaged-item or policy-conflict cases that need human review rather than an invented denial.
You receive a source-linked internal assessment, the smallest useful next step, a customer-ready draft for review, and a send check. Source links or IDs refer to the evidence you supply; the skill does not independently verify inaccessible systems. When evidence is missing, it produces a safe interim reply and targeted questions instead of inventing store rules.
Includes the core SKILL.md, a detailed decision guide, four worked examples, and reusable intake and output templates. English-first; other reply languages can be requested. Start by pasting the ticket and the relevant policy. No coding or integration is needed for that workflow.
This is a review-and-drafting skill, not an automated refund agent. It does not contact customers, create labels, issue refunds, modify orders, decide legal consumer entitlements, or guarantee business outcomes.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for E-commerce.
Draft-only: never sends messages, issues refunds, or changes orders. Does not determine legal or tax entitlements, monitor live systems, or guarantee error prevention. Results depend on the supplied evidence; missing records remain unresolved.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 7 days ago
Nothing to install. It runs in your AI agent.
Provide the customer message, relevant policy, and available order/return/refund records. Pasted-input use needs no code or integration. Optional connected sources require authorized read-only access. A human must review drafts and decide remedies.
Permissions
Reads supplied text and included guides; file access only for user-provided attachments. Drafting only: no messages sent, refunds, labels, or order changes. No credentials or integrations required. Never include payment secrets or unnecessary personal data.