The Agent Skills Directory

[SAFE]: The skill includes explicit security directives and guides for verifying Slack request signatures, validating OAuth state parameters for CSRF protection, and managing secrets using environment variables or vaults.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface due to the following factors:
Ingestion points: Slack event payloads containing untrusted user-supplied message text enter the agent context via handlers for app_mention and message events in guides/04-events-api.md and the provided app scaffolds.
Boundary markers: The provided code templates do not implement explicit boundary markers or 'ignore' instructions for processed message content.
Capability inventory: The skill enables capabilities such as posting messages to channels (say(), chat.postMessage) and opening/updating modals (views.open, views.update), which can be influenced by the content of ingested messages.
Sanitization: Examples show direct interpolation of user input (e.g., command['text'] or event['text']) into response messages and modal blocks without extensive sanitization.

slack-app-weapon