Back to skill

Security audit

city-walker

Security checks for vulnerabilities and agentic risk

Overview

This is a Chinese-language city walking route helper with disclosed POI search behavior and no executable code, persistence, or credential handling.

Install this if you want a Chinese-language assistant for walking itineraries. Be aware it may ask for your current area or starting point and use POI/search tools to make recommendations; avoid sharing precise location details if you do not need highly specific routing.

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 Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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 (5)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill description uses very broad phrases like '逛逛城市', '探索附近', and '找个地方散步', which can match many ordinary user requests and cause the skill to activate outside its intended scope. Over-broad activation can lead to unintended routing, user confusion, and inappropriate use of external tools or location-based recommendations when the user did not actually request this skill.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
78% confidence
Finding

The skill's description, triggers, examples, and output format are entirely written in Chinese, with no indication that users may choose another language or that the skill is limited to a Chinese-speaking audience for compliance reasons. Under the language/locale policy, forcing a specific language without user opt-in can be a natural-language policy violation.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger scenarios are ambiguous and include common conversational statements such as '今天想出去走走' and '附近有什么好玩的', which are not specific enough to reliably indicate a request for this exact skill. In a multi-skill environment, this increases the risk of wrongful invocation and may expose unnecessary location handling or cause the system to prioritize this skill over more appropriate ones.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill content is entirely in Chinese and does not declare a locale or language constraint, which can cause it to be surfaced to users or agents operating in other languages. That mismatch can lead to user confusion, incorrect use, or unsafe misunderstanding of route details and practical instructions, especially where place names, timing, or safety notes matter.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The skill advertises activation using broad, everyday phrases such as wanting to 'walk around the city' or 'explore nearby', without clear boundaries that distinguish it from many other location, travel, leisure, or recommendation skills. This can cause unintended invocation, leading the agent to route user requests into this skill when the user did not explicitly want a city-walking itinerary, which increases the risk of misfires and inappropriate handling of location-sensitive requests.

Content

No source excerpt is available for this finding.