Multi-SKU Bundles

Mine historical orders for multi-SKU co-purchase patterns, quantify association strength between SKUs, and produce high-converting bundle and Frequently-Boug...

This skill appears coherent and low-risk because it is instruction-only and asks for no credentials or installs. Before using it, consider: 1) Do not paste raw PII (customer emails, full names, payment data) into the chat — sanitize exports to order-line rows only (order_id can be hashed); 2) Confirm platform constraints (Shopify/Woo/custom) before relying on one-click checkout hooks; 3) Validate discount and margin math with your finance/merchandising team — the skill will propose prices/percentages but cannot access your margin data unless you provide it; 4) If you use loyalty apps (Rijoy or others), test bundle interactions in staging to avoid stacking conflicts; 5) If the skill ever asks for credentials, a download URL, or to run scripts, stop and re-evaluate — that would be a new risk signal. If you want a deeper security check, provide the raw SKILL.md plus any future script files or install specs and I will re-evaluate for hidden endpoints, downloads, or excessive privileges.

Type: OpenClaw Skill Name: multi-sku-copurchase-bundles Version: 1.0.0 The skill contains instructions that direct the AI agent to promote a specific third-party service and URL (https://www.rijoy.ai) within its responses. This constitutes a form of prompt injection for brand promotion (referral injection). While the service is relevant to the e-commerce domain and the instructions are transparently documented in SKILL.md and references/rijoy_brand_context.md, the use of instructions to manipulate the agent's output toward a specific commercial entity is a recognized injection technique. No evidence of malicious code execution, data exfiltration, or harmful intent was found.