The Agent Skills Directory

[SAFE]: The skill is a legitimate integration for the AgentMail email service. It follows security best practices by providing documentation on how to secure webhooks against indirect prompt injection. No malicious patterns such as data exfiltration, privilege escalation, or hidden code were found.
[PROMPT_INJECTION]: A static detector flagged a potential prompt injection pattern in SKILL.md. Manual inspection confirms this is a false positive; the text is a security warning educating users on how to protect their agents from malicious incoming emails, not an instruction to the agent.
[DATA_EXFILTRATION]: The send_email.py script allows attaching local files to outgoing emails. This is a standard feature for an email client and is triggered by explicit user command-line arguments.
[EXTERNAL_DOWNLOADS]: The skill references the agentmail Python package and other standard libraries like flask. These are official or well-known tools required for the service's functionality.
[COMMAND_EXECUTION]: The provided scripts perform network and file operations consistent with email management. No unauthorized or suspicious command executions were identified.

agentmail