Security audit

anniversary

Security checks across malware telemetry and agentic risk

Overview

This travel skill is mostly coherent, but it can silently keep local logs of raw travel requests and install an external CLI globally.

Review before installing. Use only if you are comfortable with a global npm-installed flyai CLI and sharing travel details with the provider. Ask the agent not to create .flyai-execution-log.json, or require redaction and deletion of raw queries before use.

SkillSpector

By NVIDIA

Vulnerability Patterns

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
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (2)

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The runbook explicitly logs `user_query` as raw input and instructs appending the full execution log to a local file. In a travel-booking skill, user input can contain names, dates, destinations, contact details, passport/visa details, or other sensitive travel context, so persistent plaintext logging creates a realistic privacy and data exposure risk if the host, logs, or downstream tooling are accessed.

Ssd 3

Medium

Confidence: 96% confidence
Finding: The schema requires persistent logging of both raw user input and detailed internal execution data such as commands, fallback actions, and output characteristics. Those records can leak sensitive user data as well as operational details that aid further compromise or misuse, especially because the log is written as newline-appended JSON to a local file with no stated protections, retention policy, or warning.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.