The Agent Skills Directory

[Data Exposure & Exfiltration] (LOW): The script scripts/send_email.py performs network operations to an external SMTP server. While necessary for the skill's purpose, this creates a channel for data to leave the environment.
[Indirect Prompt Injection] (LOW): The skill possesses a high attack surface for indirect prompt injection because it can read any file accessible to the agent and send it via email.
Ingestion points: The --html and --attachment arguments in scripts/send_email.py allow the agent to specify file paths for reading.
Boundary markers: None. There are no instructions or checks to prevent the agent from accessing sensitive files like SSH keys or environment configurations.
Capability inventory: The script uses the open() function for file reading and the smtplib library for network transmission.
Sanitization: No path validation or sanitization is implemented to restrict the tool to specific directories or file types.

email-sender