Cuihua Dependency Updater

• Outcome: Review
• Latest audit: May 29, 2026 at 9:27 PM UTC
• Version:
  1.0.0

Overview

This dependency updater is not clearly malicious, but it advertises safe automated updating, testing, rollback, and security analysis that the shipped code does not actually provide.

Review before installing. Treat this as a basic npm outdated reporter, not a safe updater. Use it only in version-controlled projects, inspect any generated package-manager commands before running them, and require explicit approval before changes to package.json, lockfiles, or node_modules.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (4)

Description-Behavior Mismatch

Medium

• Confidence: 94% confidence
• Finding: The skill markets itself as providing safety checks, breaking-change detection, and automated testing, but the implementation only runs `npm outdated --json` and uses a major-version heuristic. This can mislead users into trusting update recommendations as validated or low-risk when no real compatibility or test verification has occurred, increasing the chance of disruptive or insecure dependency changes.

Intent-Code Divergence

Low

• Confidence: 83% confidence
• Finding: The explicit 'Simple heuristic' comment confirms that breaking-change detection is not actually implemented despite the skill's claims of intelligent analysis. In the context of a dependency updater, this is dangerous because users may rely on underpowered risk classification to approve package updates that introduce regressions or incompatible API changes.

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The skill encourages automated dependency updates but does not clearly warn that running update commands can modify package manifests, lockfiles, and installed dependency trees. In an agent setting, this can lead to unintended project changes, broken builds, or silent state drift if the user does not realize the skill may make direct repository modifications.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The configuration explicitly promotes unattended auto-updates for security and patch releases without an equally prominent warning about automatic modification of project files. In context, this is more dangerous because the skill is designed to operate on dependency state, so users may enable automation that changes manifests and lockfiles across a codebase without sufficient review or rollback guarantees.

VirusTotal

65/65 vendors flagged this skill as clean.