The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documentation instructs on installing standard ecosystem packages such as sails-hook-sockets and @sailshq/socket.io-redis via npm. These resources originate from the vendor's own verified repositories and organizations.
[INDIRECT_PROMPT_INJECTION]: The skill provides instructions for building applications that ingest untrusted data from WebSocket clients, creating a potential attack surface if not properly secured.
Ingestion points: External data is received via io.socket virtual request methods and processed by server-side actions.
Boundary markers: The documentation includes security patterns such as using beforeConnect for connection-level authorization and origin whitelisting (onlyAllowOrigins) to prevent hijacking.
Capability inventory: The skill enables capabilities such as broadcasting messages to specific rooms, publishing model updates, and performing database operations based on socket inputs.
Sanitization: The provided code examples demonstrate the use of Sails.js action inputs validation to strictly define and sanitize data types, lengths, and requirements for all incoming socket payloads.

realtime