The Agent Skills Directory

[SAFE]: The skill serves as technical documentation for Windmill Triggers, providing YAML schemas for various integration types including HTTP, WebSockets, Kafka, MQTT, and Email.
[COMMAND_EXECUTION]: Contains documentation for standard CLI commands (wmill sync push and wmill sync pull) used to deploy or retrieve trigger configurations from the Windmill platform.
[INDIRECT_PROMPT_INJECTION]: Defines the attack surface for data ingestion from external sources (e.g., email bodies, webhook payloads, message queue data). While this represents a vector for untrusted data to enter the agent's context, it is the core architectural purpose of the skill and does not contain malicious patterns.
Ingestion points: HTTP requests, WebSocket messages, Kafka topics, NATS subjects, Postgres CDC, MQTT topics, AMQP queues, SQS queues, GCP Pub/Sub, Azure Event Grid, and incoming emails.
Boundary markers: No explicit sanitization requirements are defined in the schemas; input validation is handled by the invoked scripts/flows.
Capability inventory: Scripts triggered by these events have access to the wmill SDK for interacting with platform resources like S3 storage.
Sanitization: Not explicitly implemented in the provided configuration schemas.

triggers