Security audit

Multi-SKU Bundles

Security checks for vulnerabilities and agentic risk

Overview

This is a benign instruction-only helper for turning order history into bundle recommendations, with a reminder to share only necessary order data.

Safe to install as a bundle-analysis helper. When using it, remove customer names, emails, addresses, payment details, and other unnecessary personal data from order exports before sharing them.

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (2)

Vague Triggers

Medium

Confidence: 89% confidence
Finding: The skill’s trigger description is unusually broad and includes common business phrases like raising AOV, bundle design, Shopify bundle apps, and even partial phrases such as 'customers who buy this also buy'. This can cause the skill to activate in conversations that are only loosely related, increasing the chance of unintended routing, inappropriate data handling expectations, or the model applying this skill’s rigid output format when a different skill would be safer or more accurate.

Vague Triggers

Low

Confidence: 91% confidence
Finding: The eval encourages activation on very broad phrases like bundle advice, cross-sell, or 'customers who buy this also buy,' which can cause the skill to trigger on ordinary merchandising requests that may not actually require basket-analysis logic. This is not a code-execution issue, but it can lead to over-broad routing, irrelevant outputs, and reduced reliability by making the agent apply specialized behavior outside its intended scope.

Static analysis

No suspicious patterns detected.