You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Apr 28, 2026. It is now read-only.
Ad-hoc agent messaging causes inconsistent behavior and higher latency. There’s no standardized envelope or lightweight broker for topic-based fanout and direct messages between agents.
Describe the solution you'd like
Introduce a compact Message/MessageType protocol and an async in-process A2ABroker supporting:
Topic pub/sub for events and coordination
Direct send to agent inboxes
Drop-on-full queues to keep common-path latency <100ms
Describe alternatives you've considered
External brokers (NATS/Redis/Kafka): robust but add ops overhead and dependencies.
HTTP between agents: simple, but adds per-request connection costs and lacks easy fanout.
Use Case
Agents coordinating cluster actions (scale, rollout, health events) with low latency and simple developer ergonomics. For example, broadcasting “wec-east1 ready” to subscribers or sending a direct “ping” to a coordinator.
Priority
High - Important for my workflow
Feature Type
New functionality
Performance improvement
User experience enhancement
Developer experience improvement
Security enhancement
Documentation improvement
API enhancement
Additional Context
See short design overview in docs/a2a_mcp_design.md. Broker is intentionally minimal to be replaced by a distributed backend later without changing call sites.
Implementation
I would be willing to submit a pull request to implement this feature
Contact Details
alokdangre@gmail.com
Is your feature request related to a problem?
Ad-hoc agent messaging causes inconsistent behavior and higher latency. There’s no standardized envelope or lightweight broker for topic-based fanout and direct messages between agents.
Describe the solution you'd like
Introduce a compact
Message/MessageTypeprotocol and an async in-processA2ABrokersupporting:Describe alternatives you've considered
Use Case
Agents coordinating cluster actions (scale, rollout, health events) with low latency and simple developer ergonomics. For example, broadcasting “wec-east1 ready” to subscribers or sending a direct “ping” to a coordinator.
Priority
High - Important for my workflow
Feature Type
Additional Context
See short design overview in
docs/a2a_mcp_design.md. Broker is intentionally minimal to be replaced by a distributed backend later without changing call sites.Implementation
Code of Conduct