The Agent Skills Directory

COMMAND_EXECUTION (MEDIUM): The skill utilizes the vercel CLI to perform operations. While these are legitimate infrastructure management tasks, they involve subprocess calls that execute system-level commands.
EXTERNAL_DOWNLOADS (LOW): The skill references the installation of the vercel CLI via npm or brew. Per [TRUST-SCOPE-RULE], Vercel is a trusted organization, downgrading this finding to LOW as it points to official, reputable software.
PROMPT_INJECTION (HIGH): This skill is highly susceptible to Indirect Prompt Injection (Category 8). It possesses high-privilege write/execute capabilities (modifying DNS records) and accepts external parameters like domain names and TXT record values.
Ingestion points: Command-line arguments for <domain>, <subdomain>, <target>, and <value> provided in the documentation.
Boundary markers: None detected. There are no instructions to the agent to treat these inputs as untrusted or to ignore embedded instructions within processed data.
Capability inventory: The skill can create (add) and delete (rm) DNS records. This allows for traffic redirection, MX record hijacking for email interception, and ownership verification for malicious services.
Sanitization: No sanitization or validation logic is present in the skill definition to ensure inputs are valid domain components or to prevent command injection through shell meta-characters (though the Vercel CLI itself may provide some protection).

vercel-dns