from drift771
Provide robust server-sent-events (SSE) stream management with Redis-backed registry, heartbeat monitoring, completion persistence, and background guardian clea
This skill adds resilient SSE stream management to a web backend. It provides a Redis-backed stream registry to track active streams, periodic heartbeat updates to detect stale or orphaned connections, a small completion store to persist terminal events for client recovery, and a background guardian process that cleans up abandoned streams. Together these pieces reduce lost terminal events, prevent resource leaks, and make client reconnection deterministic.
Use this when your application exposes long-lived SSE connections for real-time updates (progress, streaming generations, logs) and you need: reliable recovery when clients disconnect and reconnect, detection of silent disconnects, and automatic cleanup of abandoned server-side resources. It's suitable for generators, progress streams, live logs, or any system where clients may drop and later reconnect.
Best fit for TypeScript/Node.js agents and developer workflows that can run server-side TypeScript. Integrates naturally with backend frameworks that can use Redis (Express/Next.js/Cloud functions).
SSE Stream Resilience provides Redis-backed SSE stream management with heartbeat monitoring, completion persistence, and a background guardian for orphan cleanup. The SKILL.md is essentially a code tutorial with full TypeScript implementations but contains no executable scripts — it's a reference pattern, not an actionable agent skill. Code is well-structured with clear types and Redis pipeline usage, but the skill doesn't guide an agent to perform specific tasks; it's more of a developer reference document.
This is a code tutorial/reference rather than a true agent skill. It provides solid TypeScript implementations for SSE resilience patterns but doesn't define triggers, agent actions, or workflow steps. The frontmatter is minimal (missing description field consistency). Security is decent — no shell injection, no credential leaks, no destructive commands. The empty catch blocks in the SSE endpoint and client code are a minor concern. The 'depreciated' in the source path suggests this may be stale content.
