vultr

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The prompt includes examples that place the Vultr API key directly on the command line (e.g., vultr-cli config set api-key YOUR_API_KEY_HERE) and shows exporting the key in plaintext, meaning an agent that automates these steps or echoes commands would need to handle and potentially output the secret verbatim—an exfiltration risk.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill's prerequisites instruct installing the vultr-cli by fetching remote executables (e.g. curl -sL https://github.com/vultr/vultr-cli/releases/latest/download/vultr-cli_linux_amd64.tar.gz and other git/go install references like https://github.com/vultr/scoop-vultr.git and github.com/vultr/vultr-cli/v3@latest), which are required runtime dependencies that download remote code that may be executed on the host.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.90). The skill instructs the agent to run state-changing commands on the host (e.g., ssh-keygen that writes to ~/.ssh, echo >> .env.local, and a sudo mv to /usr/local/bin) and therefore pushes the agent to obtain/use elevated privileges and modify system files.