The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads the vultr-cli tool directly from the vendor's official GitHub repository during the installation process.
[REMOTE_CODE_EXECUTION]: As part of the prerequisite setup, the skill downloads a binary archive and extracts it for execution to enable cloud management capabilities.
[COMMAND_EXECUTION]: The skill uses vultr-cli for server provisioning, firewall configuration, and Kubernetes cluster management, alongside ssh for remote server verification.
[DATA_EXFILTRATION]: The skill accesses the user's public SSH key for deployment purposes and stores configuration variables, including IP addresses and local key paths, in a .env.local file.
[PRIVILEGE_ESCALATION]: The installation instructions utilize sudo to move the vultr-cli binary to a system directory (/usr/local/bin/), which is a standard procedure for installing command-line tools.
[DYNAMIC_EXECUTION]: The skill generates local environment configuration scripts at runtime by appending deployment metadata to a local file.
[INDIRECT_PROMPT_INJECTION]: The skill interpolates user-supplied parameters into shell commands, representing a potential surface for injection if inputs are not validated by the agent platform.
Ingestion points: SERVER_NAME and SSH_KEY_NAME parameters collected in SKILL.md and assets/env-template.
Boundary markers: No specific delimiters or safety instructions are used when interpolating these values into commands like grep or vultr-cli instance create.
Capability inventory: The skill executes vultr-cli, ssh, curl, and shell processing tools (grep, awk, echo).
Sanitization: No sanitization or escaping mechanisms are applied to user-provided strings before they are executed in a shell context.

vultr