Back to skill

Security audit

Multi Search CN

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed Chinese web-search helper that sends search queries to DuckDuckGo and provides links to Chinese search engines, with no evidence of persistence, credential access, or destructive behavior.

Install this if you want Chinese-locale search assistance. Avoid using it for sensitive or private queries unless you are comfortable sending those terms to DuckDuckGo and potentially opening links to Chinese search providers; use urls-only mode when you only want generated search links.

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
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding

The skill invokes a local Python script that performs web searches, but the manifest does not declare any tool scope, permissions, or allowed tools. This creates an authorization gap where an agent may use network-capable behavior without an explicit contract, reducing reviewability and increasing the chance of unintended external requests or data exposure.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The skill is explicitly China-focused by default and routes searches toward cn-zh/DDG and Chinese search engines without requiring explicit user consent. This can expose user queries to region-specific providers and alter result sets based on jurisdiction, censorship, or privacy differences, especially when the user did not ask for a China-locale search.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The activation text is broad enough to trigger on common search-related requests, including general web lookup scenarios where a more appropriate or safer search skill may exist. Overbroad routing can cause unexpected network access, accidental use of China-specific search pathways, and lower user awareness that their query will be sent to third-party engines.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The request header hard-codes Accept-Language: zh-CN,zh;q=0.9,en;q=0.8, which forces a specific locale preference for all outbound requests. This is a natural-language policy concern because the user is not given any choice or opt-in regarding language/locale behavior.

Content

No source excerpt is available for this finding.