ClawHub
Back to skill

Security audit

yoga-retreat

Security checks across malware telemetry and agentic risk

Overview

This is a flight-search skill that uses an external travel CLI, with some overbroad wording and install-risk notes but no evidence of hidden, destructive, or data-stealing behavior.

Install only if you are comfortable with the agent running flyai travel-search commands and, if missing, installing a global npm CLI. Treat it as a flight-search helper, not full-service travel booking, and review booking links before entering personal, passport, or payment information.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The skill explicitly says to never invent CLI parameters unless they appear in the documented Parameters table, but later instructs use of `--journey-type 1` for direct flights even though that flag is absent from the table. This inconsistency can cause agents to rely on undocumented behavior, fail unpredictably, or be manipulated into accepting hidden parameters as legitimate.

Vague Triggers

Medium
Confidence
79% confidence
Finding
The activation trigger includes the Chinese phrase `出行预订` ('travel booking'), which is far broader than yoga-retreat flight searches and can match ordinary travel requests unrelated to the skill's niche. Overbroad activation increases the chance the agent invokes this skill in the wrong context, causing unnecessary command execution, package installation attempts, or incorrect tool-mediated responses.

Vague Triggers

Medium
Confidence
83% confidence
Finding
The description advertises a wide range of services such as hotels, trains, attraction tickets, visas, insurance, and car rentals, while the body of the skill is centered on flight search. This mismatch broadens perceived scope and can lead orchestration systems or users to invoke the skill for unsupported tasks, increasing erroneous execution and scope confusion.

Missing User Warnings

Low
Confidence
93% confidence
Finding
The file instructs the operator to run a global npm installation command without any warning that it modifies the host system and pulls executable code from an external package source. In an agent-skill context, this is risky because recovery instructions may be followed automatically or by a trusting user, expanding impact beyond simple documentation into unintended environment modification and supply-chain exposure.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.