cRSS enables Agents to have shared memory (Agent B knows your preferences because you told Agent A) or to share behavior. (Agent B adopts behavioral characteristics of Agent A).

Not a library, framework, nor storage system.

cRSS is a primitive.

cRSS turns LLM context mutation into structured, deterministic, auditable event feed (like RSS + event sourcing + git-like diffs), so multiple agents can share, replay, audit, and deterministically reconstruct what went into each model inference.