new·The score now tells you which way it movedA brain's exam only ever grows: its own material writes questions, and so does every question a real caller asked and did not get answered. The score is a percentage over that growing set, so a brain that learned more could post a smaller number — and this week three did. One of them answered two MORE questions than the week before and showed eighteen points less. Printed as a single percentage, that reads as decline to a reader and as punishment to anyone who contributes material.all news →
mozg.beta
Sign in

LangChain & LangGraph · all subjects

agents & orchestration

4 notes, read out of this brain and free to use. Each one was extracted from a source and is re-checked against its exam.

A2A (Agent-to-Agent) protocol integration

The Agent Server implements A2A protocol support for agent interoperability. The protocol was updated to v1.0 RC with renamed JSON-RPC methods, ListTasks handler, and enhanced role, state, and part formats. A2A messageId is mapped to LangChain message IDs for proper tracking. The `/a2a/{assistant_id}/` path serves `.well-known/agent-card.json` for agent discovery. A2A maps 'interrupted' status to 'input-required' for task status alignment.

Cron scheduling and persistent cron threads

The Agent Server supports scheduled cron jobs with an `is_enabled` field to pause/resume execution. Crons can be configured with `on_run_completed: {keep,delete}` to manage thread persistence after cron execution. Crons support metadata propagation and merging on PATCH requests. Authentication for cron creation prevents privilege escalation with independent filtering for crons, threads, and assistants.

Interrupted runs metadata

Interrupted runs carry the same fields as root runs. The run type itself signals the abnormal termination state; no additional required fields are added.

Five run types in coding agent metadata schema

The schema distinguishes five run types: 'root' (top-level run representing a full agent turn or session), 'llm' (language model call within a turn), 'tool' (tool invocation within a turn), 'subagent' (nested or delegated agent run), and 'interrupted' (run interrupted before completion).

Give your agent this brain