Security audit

AI Customer Service Automation

Security checks for vulnerabilities and agentic risk

Overview

This customer-service automation skill is coherent, but it advertises automatic handling, reuse, and third-party forwarding of customer conversations without enough privacy controls or user approval guidance.

Review before installing in any real support environment. Enable only approved channels, disable or gate automatic ticket creation for sensitive cases, require redaction of PII and account data, and define retention, deletion, and access controls for historical conversation learning and exported tickets.

A.I.G

Vulnerability Patterns

Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (5)

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: Automatic ticket creation can forward user conversation details into external systems such as Jira, Feishu, or DingTalk, but the documentation does not disclose that transfer or its privacy implications. This is risky because tickets frequently persist for long periods, broaden internal access, and may replicate sensitive customer content beyond the original support channel.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The multi-channel support claim covers third-party platforms such as Feishu, WeChat, Telegram, and email but does not warn that customer messages may be transmitted to and stored by those services. Because customer support conversations often include sensitive identifiers and business information, omission of these transfer implications can lead to unsafe deployment decisions and unexpected third-party data exposure.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The documentation states that the system will automatically learn from historical conversations, which implies ingestion and reuse of potentially sensitive customer data without any disclosure of consent, retention, redaction, or access controls. In a customer service context, historical chats commonly contain personal, financial, or account information, so silent reuse materially increases privacy and compliance risk.

Natural-Language Policy Violations

Medium

Confidence: 88% confidence
Finding: The natural-language description forces a specific language presentation in the manifest and does not indicate that users can choose their preferred language or that the skill is intended only for a Chinese-speaking context. This can violate language/locale policy when the skill is distributed in a broader environment without explicit opt-in or justification.

Natural-Language Policy Violations

Low

Confidence: 87% confidence
Finding: Nearly all user-facing documentation and instructions are in Chinese, but the file does not state that the skill is region-specific or provide an opt-in language choice. This can violate language/locale policy when a skill effectively forces a language without user choice.

Static analysis

No suspicious patterns detected.