The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by processing untrusted data from incoming WhatsApp messages.
Ingestion points: WhatsApp messages are retrieved via mcp__whatsapp__list_messages in SKILL.md.
Boundary markers: There are no explicit delimiters or instructions to the agent to ignore potentially malicious content within the message body during parsing.
Capability inventory: The skill can send messages via mcp__whatsapp__send_message, download media via mcp__whatsapp__download_media, and create tasks/notes in the system.
Sanitization: The provided parseQuickCapture logic lacks sanitization or validation of the message content before processing and confirmation.
[EXTERNAL_DOWNLOADS]: The technical implementation notes reference the use of external Node.js dependencies.
Evidence: The code snippets in SKILL.md show the use of require('node-cron') for scheduling briefings.
[COMMAND_EXECUTION]: The skill documentation includes examples of command-line operations to verify connectivity.
Evidence: SKILL.md suggests running claude mcp list | grep whatsapp to check server availability.

whatsapp-message-management