Security audit

Explore Dubai

Security checks for vulnerabilities and agentic risk

Overview

This Dubai travel skill is mostly coherent, but it can install an unpinned global CLI and quietly retain raw travel prompts in a local log file.

Install only if you are comfortable with a travel skill that may install and run a provider-specific global npm CLI. Avoid putting passport numbers, payment details, booking references, or other sensitive personal data into prompts, and check for or delete `.flyai-execution-log.json` if you do not want local query logs retained. Verify visa rules with official government sources.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (4)

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill mandates installing and running a global npm package and querying an external CLI without any trust, provenance, permission, or data-handling warning. In an agent context, this can cause unreviewed system modification and outbound data disclosure, especially if user queries or environment-derived values are passed to the tool automatically.

Missing User Warnings

Medium

Confidence: 97% confidence
Finding: The runbook explicitly logs `user_query` as raw input in an internal execution log, creating a persistent record of whatever the user typed. In a travel skill, users may include passport details, contact information, booking identifiers, dates of birth, or payment-adjacent data, so storing raw prompts without minimization or user notice increases privacy and data-exposure risk.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The runbook directs the agent to append execution logs to a local file, but provides no controls for retention, access restriction, encryption, or deletion. Persistent file-based logs can be harvested by other processes, exposed through backups, or retained indefinitely, turning routine operational metadata into a long-lived privacy and security liability.

Ssd 3

Medium

Confidence: 98% confidence
Finding: Taken together, the runbook both captures the user's raw input and persists the resulting execution log to disk, which materially increases the chance that sensitive user-provided travel data will be retained and later disclosed. Because this skill supports bookings, visas, insurance, and reservations, user prompts are especially likely to contain personal and financial-context information, making the combined logging behavior more dangerous than in a low-sensitivity skill.

Static analysis

No suspicious patterns detected.