AirDrop

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:19 PM UTC
• Version:
  1.0.0

Overview

This skill is a local macOS AirDrop helper that shares user-selected files and does not show hidden upload, credential access, or destructive behavior.

Install only if you are comfortable with a local helper opening macOS AirDrop for exact files you choose. Confirm the payload before sharing, avoid secrets and hidden files unless intentionally included, and delete ~/airdrop/staging files after use if they contain sensitive content.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

Vague Triggers

Medium

• Confidence: 88% confidence
• Finding: The activation rule is intentionally broad, causing the skill to trigger on common requests about sending files to Apple devices rather than only explicit AirDrop intent. In an agent system, this can lead to over-activation, unnecessary memory writes about user behavior, and unintended invocation in adjacent contexts involving local file handling or device transfers.

Missing User Warnings

Low

• Confidence: 86% confidence
• Finding: The recipe instructs the agent to create directories, copy files into a staging area, and package them into a ZIP without explicitly warning that this performs additional local disk writes and creates duplicate copies of potentially sensitive files. In an automation context, that can leave residual artifacts in predictable locations and expand the set of files exposed to local users, backups, or later accidental sharing.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: This workflow tells the agent to write conversational summary text to a file and then share it, but it does not warn that chat content will be persisted to disk first. That creates a confidentiality risk because transient conversation data may contain secrets, personal data, or internal notes that become recoverable from the filesystem, backups, or later reuse.

VirusTotal

66/66 vendors flagged this skill as clean.