The Agent Skills Directory

[SAFE]: The skill follows standard architectural patterns for WebSocket implementations, including horizontal scaling with Redis and sticky sessions. No evidence of obfuscation, credential harvesting, or unauthorized network operations was found.
[INDIRECT_PROMPT_INJECTION]: The skill defines patterns for processing untrusted data received from WebSocket clients, creating a theoretical attack surface for indirect injection.
Ingestion points: Message handlers such as socket.on("message", ...) in SKILL.md and multiple event listeners throughout the reference files.
Boundary markers: Absent, as the skill focuses on technical implementation templates rather than natural language boundary enforcement.
Capability inventory: The provided code includes the ability to broadcast messages to rooms, manage user presence in Redis, and interact with session cookies.
Sanitization: The skill explicitly includes a security reference guide (references/security.md) that demonstrates mitigation strategies, including input validation using Joi and HTML sanitization using sanitize-html to prevent XSS and other injection attacks.

websocket-engineer