01 / The short version
What happened
LangSmith released Managed Deep Agents v0.9 in public beta, introducing self-scheduling via a new Schedules SDK, per-run dynamic configuration, and Slack reactions. Agents can now create recurring reminders and follow-ups that run with the requester's permissions. The dynamic configuration allows a single deployment to load team-specific skills, models, and tools based on runtime context, improving isolation and reducing context size. Slack reactions provide immediate visual feedback, customizable per message or channel.
See the exact references02 / Key takeaways
What you need to know
- 01
Agents can autonomously schedule reminders and recurring tasks using cron expressions or specific times.
- 02
Per-run configuration allows dynamic selection of models, skills, and tools based on channel or user context.
- 03
Slack reactions are enabled by default to signal agent activity and can be customized.
Keep in perspective
What to watch for
Verify if the public beta access requires specific LangSmith tier subscriptions or if there are limitations on the number of concurrent scheduled tasks.
Go to the source
Exact references
These are the original pages used for this brief. Publisher claims are not independent evaluations.
01Primary source · LangChainRead the original announcementhttps://www.langchain.com/blog/managed-deep-agents-schedules-per-run-configuration-slackAI-generated from the linked source. It can miss context; verify consequential details in the original. How the radar works