Back to skill

Security audit

Travel Concierge CLI

Security checks for vulnerabilities and agentic risk

Overview

The skill is not overtly malicious, but it asks agents to run an unbundled contact-scraping CLI with broad triggers and unclear provenance.

Review this skill before installing. It should only be used for user-provided accommodation listing URLs, and the `travel-concierge` executable should come from a known, trusted source before an agent runs it. Consider narrowing triggers and clarifying API key storage and scraping limits.

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 (4)

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The declared purpose is to retrieve accommodation contact information, but the supplied code does not implement any contact-finding logic, platform interaction, scraping, API usage, or data retrieval. Instead, it only configures the Vitest test runner and coverage settings. This is a materially different primary purpose and therefore a clear description-behavior mismatch.

Content

No source excerpt is available for this finding.

Shadow Command Trigger

Medium
Category
Trigger Abuse
Confidence
82% confidence
Finding

The trigger phrase "find contact" begins with a common verb that may conflict with or shadow a built-in "find" command pattern, creating ambiguous routing and increasing the chance this skill intercepts generic search requests. In agent systems, shadowing built-ins can produce surprising tool execution and widen exposure to unintended data collection actions.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger phrase "find contact" is very broad and can activate on many unrelated user requests, causing this skill to run outside its intended accommodation-listing context. In an agent ecosystem, overly broad triggers can lead to unintended scraping behavior, privacy-invasive contact collection, or execution in situations where the user did not intend to use this tool.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

Phrases like "property contact," "direct booking," "property email," and "property phone" are ambiguous and could match many real-estate, travel, or generic business-contact requests. Because the skill is designed to gather off-platform contact methods for listings, accidental invocation expands the chance of collecting or surfacing personal/business contact data in contexts the user did not clearly request.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.