Back to skill

Security audit

application-use

Security checks for vulnerabilities and agentic risk

Overview

The skill is a disclosed macOS automation tool, but it gives agents broad GUI control with little safety scoping and recommends an unpinned global npm install.

Review carefully before installing. Use only if you are comfortable giving an agent broad control over macOS applications, visible screen contents, focused text fields, browser sessions, and Finder actions. Prefer a pinned reviewed package version, avoid entering passwords or approving financial/account actions through automation, and require explicit confirmation before submissions, file changes, screenshots, or interactions with sensitive apps.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:13
Finding

Unpinned Global npm Package Installation

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, line 13
Vulnerability Type: Supply-chain exposure through an unpinned third-party dependency
Risk Level: Medium

Vulnerable Code:

bash
npm i -g application-use

Technical Analysis

The installation instruction retrieves the current npm registry version of application-use without specifying an exact version, lockfile, or integrity hash. Consequently, the installed code may differ from the version originally reviewed.

npm packages can execute lifecycle scripts during installation. If the package, its publisher account, or a transitive dependency is compromised, following this instruction could execute attacker-controlled code with the privileges of the user running npm. Installing the package globally also makes its executable available outside this project and increases the scope and lifetime of a compromised installation.

This finding identifies unsafe dependency acquisition rather than confirmed malicious behavior in the named package. The audited project contains no embedded payload demonstrating that application-use is currently malicious.

Attack Path

  1. An attacker compromises the package publisher, the npm package, or an included dependency, or otherwise causes a malicious release to become the version resolved by npm.
  2. A user follows the installation command in SKILL.md.
  3. npm downloads the mutable registry version rather than a previously reviewed, fixed artifact.
  4. Malicious package code or an installation lifecycle script executes under the invoking user's account.
  5. The globally installed command remains available and may execute again during later desktop-automation tasks.

Impact Assessment

Successful exploitation could permit arbitrary code execution with the invoking user's privileges. Depending on those privileges and host protections, an attacker could access user-readable files, environment variables, appl ...[truncated 262 chars]

Remediation
View remediation

Remediation Suggestions

  • Pin the package to an exact, reviewed version rather than resolving the latest release, for example application-use@X.Y.Z.
  • Prefer a project-local dependency recorded in package.json and a committed lockfile instead of a global installation.
  • Use npm ci with a reviewed lockfile in controlled deployment workflows.
  • Verify package provenance, publisher identity, release signatures or attestations, and registry integrity metadata before installation.
  • Audit the selected package version and its transitive dependencies for lifecycle scripts and known vulnerabilities.
  • Disable lifecycle scripts with --ignore-scripts where the package can function without them.
  • Run the tool under a least-privileged account and limit its access to sensitive files, credentials, applications, and network destinations.
  • Document the official package source and approved version-update procedure so upgrades require security review.
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger description uses very broad natural-language phrases such as "open an app," "click a button," and "fill a form," which are common requests in many benign conversations. This can cause the skill to be invoked in contexts the user did not explicitly intend, increasing the chance that a high-privilege desktop automation tool performs sensitive GUI actions on the user's machine.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The skill enables powerful macOS desktop automation capabilities, including opening apps, typing into focused elements, navigating browsers, taking screenshots, and interacting with files, but it provides no safety guidance about privacy, destructive actions, or user-consent boundaries. In this context, the omission is dangerous because GUI automation can expose sensitive data, submit unintended actions, or alter local state through clicks and keystrokes that affect arbitrary applications.

Content

No source excerpt is available for this finding.