The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection where malicious instructions could be embedded in meeting data.\n
Ingestion points: Untrusted data enters the skill via the meetingSubject and meetingTime parameters in the main.mo file.\n
Boundary markers: The skill does not implement delimiters or provide instructions for the agent to ignore embedded commands when interpolating variables into the email body content.\n
Capability inventory: The skill uses the EmailClient.sendRawEmail function (defined in emailClient.mo) which performs network operations to transmit data externally via email.\n
Sanitization: There is no evidence of input validation, escaping, or sanitization for the data processed before it is included in the final email output.

email-raw