SDK Architecture Guides
Implementation guidance for SLOP SDK developers. These documents cover how to build providers and consumers — the internal architecture of the SDKs, not the wire protocol.
The protocol spec defines what flows over the wire. These guides define how SDKs should build, serve, and debug state trees internally.
Documents
Section titled “Documents”-
Discovery & Bridge — Provider scanning, extension bridge client/server, relay transport, auto-connect, state formatting. Language-agnostic specification for the discovery layer that sits above the consumer SDK.
-
Development & Debugging —
printTree(), wire-format inspection, schema validation, message logging, common debugging scenarios. -
Sessions & Multi-User — How SDKs handle multi-user server apps. Covers session-scoped trees vs provider-per-session, scaling tradeoffs, session-aware descriptors, scoped refresh, and meta-framework adapter patterns.
Related spec content
Section titled “Related spec content”Some spec documents contain SDK-specific sections alongside protocol definitions:
-
Web Integration — Package architecture (layers 0–4),
createSlop()/createSlopServer()API, descriptor format,useSlop()hooks, typed schema, scoped clients, framework adapters, transport adapters, meta-framework helpers. -
Scaling — The “Developer API for scaling” section covers
slop.register()patterns for summaries, windowed collections, depth control, and salience in descriptors.
These remain in the spec because they’re tightly interleaved with protocol concepts. The SDK-specific sections are clearly marked.