OpenClaw to check books and fill a Taobao cart
A parent writing as Daretobeyourself M Y F gradually extended an OpenClaw shopping-list agent into a picture-book buying workflow. Gemini first researched recommendations from the child's age and interests. OpenClaw then checked editions, prices and other parents' reviews before adding selected books to Taobao.
medium.com
Original prompt
Shared by the source author. See source notes for context.
just add to cart — I’ll handle payment myself
Reach people exploring AI agents
See placements →What happened
A verified picture-book shortlist organized into buying batches and a Taobao cart used for the parent's purchase.
Reported by Daretobeyourself M Y F (@daretobeyourself.m.y.f)How the agent did it
- Research age-appropriate candidates, then verify editions and parent discussions.
- Require source-post links and inspect comments rather than only likes.
- Populate the Taobao cart for a human phone review and payment.
Where a person stepped in
The parent caught fabricated research, strengthened evidence rules, inspected the cart and paid.
A parent writing as Daretobeyourself M Y F gradually extended an OpenClaw shopping-list agent into a picture-book buying workflow. Gemini first researched recommendations from the child's age and interests. OpenClaw then checked editions, prices and other parents' reviews before adding selected books to Taobao.
The review step needed repair. An earlier recommendation had sounded researched but proved fabricated when the author required the agent to open Xiaohongshu and search. They developed a browser-based search skill through repeated debugging, using Brave's CDP connection. Recommendations eventually had to include full post links and evidence.
The author also changed search sorting from likes to comment count, looking for substantive parent experiences and complaints rather than relying on popular promotional posts. For the picture books, that verification produced batches of what to buy first and what could wait.
A separate Taobao browser skill searched the web store, compared product specifications and prices, and added the chosen books to the cart. The author opened the cart on a phone, checked the prices and paid. This is an actual purchase account with human checkout, not an autonomous payment claim. The book titles, child's age and total cost are not published. The useful pattern is to verify recommendations against linked discussions and exact editions before delegating cart entry.
Prompt
Original prompt
Source: medium.com · Daretobeyourself M Y F (@daretobeyourself.m.y.f)
Frequently asked questions
Why change the review search?
An earlier recommendation was fabricated, so the author required actual post links and stronger evidence.
Why sort by comments?
The author found parent experiences and complaints more useful than popular promotional posts.
Who completed payment?
The parent checked the cart on a phone and paid.
More details
About this account
OpenClaw searched reviews and products, organized buying batches and populated the cart.
Firsthand primary article opened and read. Prompt is the author's verbatim reported instruction limiting the agent to cart entry; it is not a full picture-book research prompt.
Credit: Daretobeyourself M Y F (@daretobeyourself.m.y.f)
Source published 7 May 2026
Source checked 14 Oct 2026
Tools and topics
Brave CDP · Xiaohongshu · Taobao · Gemini
Picture books · Review verification · Taobao cart
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.