Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Browses one or more shopping sites for a described item, collects candidates with price and review data, and handles checkout using secure-autofill for credential filling. Browsing runs in a sub-agent so the main chat stays unblocked. Results are saved to a local JSON artifact and the user must accept or deny candidates before any checkout proceeds.
Combining multi-site browsing, structured candidate comparison, and credential-safe checkout in one flow removes the need to switch between browser tabs, spreadsheets, and a password manager.
When to use it
- Finding the lowest price for a specific product across Amazon and Walmart
- Comparing review scores for electronics before committing to a purchase
- Placing an order without manually entering payment or shipping details
- Shopping with a hard budget cap across multiple retailers at once
- Researching shipping speed differences between sites before buying
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Find me a standing desk mat under $60, prioritizing reviews
- 1Spawns a sub-agent to browse configured sites starting from search result URLs
- 2Extracts up to maxCandidatesPerSite items per site using compact snapshots to avoid context overflow
- 3Records each candidate with price, review score, URL, and a short verdict to a workspace artifact JSON
- 4Sets artifact phase to awaiting_accept_deny and presents a ranked summary to the user
- 5After the user accepts a candidate, confirms checkout intent, then uses secure-autofill to fill credentials and submit the order
Order placed for the accepted candidate; artifact updated with status shopped and phase ordered
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.