ClawHub
Back to skill

Security audit

Wundervault Vault

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed credential-vault integration with high-impact secret and command capabilities that fit its stated purpose, but users should handle setup links and stored credentials carefully.

Install only if you trust Wundervault and need an agent to use secrets for commands, deploys, or config files. Treat the setup URL and generated credential files as secrets, verify the npm package and onboarding script checksums before use, grant each agent only the entries it needs, keep high-impact keys at an approval tier, and enable .env injection only for projects where persistent on-disk secrets are acceptable.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Intent-Code Divergence

Medium
Confidence
80% confidence
Finding
The documentation overstates the safety property by saying plaintext secrets are never returned to the agent, while later features explicitly allow secrets to be injected into commands and written to .env files on disk. Even if plaintext is not printed in chat, placing secrets into an agent-controlled process environment or file system materially exposes them to code the agent can influence, which can mislead users about the real trust boundary.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The instructions direct users to run an onboarding flow using a setup URL that contains a passphrase in the URL fragment, then store decrypted credentials in a local file and modify agent configuration, but they do not explicitly warn that the setup URL and resulting credential files are highly sensitive secrets. In a credential-management skill, this omission increases the chance users will paste the URL into logs, shell history, screenshots, or chats, or leave credential files with overly broad permissions, leading to credential exposure and downstream compromise.

VirusTotal

59/59 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.