ClawHub
Back to skill

Security audit

one-way

Security checks for vulnerabilities and agentic risk

Overview

This is a coherent travel-planning skill, but it needs review because it appears to store raw travel requests and command details in a persistent local log.

Install only if you are comfortable using a global FlyAI CLI that sends travel searches to an external travel service. Before using it for sensitive trips, confirm whether execution logs can be disabled or deleted, and avoid entering passport, visa, payment, or highly personal itinerary details unless logging and retention are clearly controlled.

SkillSpector

Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
88% confidence
Finding
The fallback in Playbook D broadens a narrowly scoped one-way flight skill into a generic keyword search workflow. That scope expansion can cause the agent to execute a less constrained command path using user-derived free text, increasing the chance of irrelevant actions, policy bypass, or unsafe downstream behavior compared with the structured flight-search commands used elsewhere.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The runbook explicitly records `user_query` as raw input in an internal execution log and then persists the log to disk via append redirection. In a travel-booking skill, raw queries commonly contain personal data such as names, locations, dates, passport or visa details, and other sensitive itinerary information, so storing them without minimization, notice, or retention controls creates a real privacy and data-exposure risk.

Ssd 3

Medium
Confidence
98% confidence
Finding
The file instructs the agent to persist an internal execution log containing raw user input, CLI commands, and operational details to `.flyai-execution-log.json`. For this skill context, users may submit highly sensitive travel and identity-related information, so persistent storage increases the blast radius of accidental disclosure, misuse, or compromise beyond the immediate request.

Static analysis

No suspicious patterns detected.