The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes the resend CLI tool to send emails directly from the shell.
[DATA_EXFILTRATION]: The skill facilitates the transmission of local data to external recipients via the Resend API.
Evidence: The --attachment and --html-file flags in SKILL.md allow the agent to read arbitrary files from the local filesystem and include them in outbound emails.
Destination: Data is sent to external email addresses provided via the --to, --cc, and --bcc flags.
[PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection attacks where malicious data can influence agent behavior.
Ingestion points: The agent takes input from external sources to populate flags such as --subject, --text, --html, and --attachment in SKILL.md.
Boundary markers: Absent. There are no instructions or delimiters defined to prevent the agent from following instructions embedded within the data it is processing.
Capability inventory: The skill combines file-read capabilities (via attachments) with network transmission (via email) using the resend emails send command.
Sanitization: Absent. Untrusted input can be passed directly to the command-line flags, which could lead to unauthorized file access if the agent is coerced into attaching sensitive files like ~/.ssh/id_rsa or .env.

resend-email