fix(langchain/createAgent): add graph support for jumps that are statically typed

[christian-bromann]

This patch updates the jump behavior of the agent, forcing user to statically type valid targets and allows us to better draw graphs from it. This introduces 2 new options to the middleware: beforeModelJumpTo and afterModelJumpTo which are lists and can contain jump targets: model, end and tools. Only if the jump target is defined the beforeModel or afterModel hook is allowed to jump to it.

I created a test to verify all combination of what user can provide:

• 2 agents
  -beforeModelJumpTo and afterModelJumpTo set to undefined, [] or either combination of the allowed values
• a tool is set or not

You can review the results and graphs in this gist: https://gist.github.com/christian-bromann/8a167bff41925364eda9c15807780b2b

[changeset-bot]

⚠️ No Changeset found

Latest commit: c78b796

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Preview	Comments	Updated (UTC)
langchainjs-docs	Ready	Preview	Comment	Sep 25, 2025 6:02am

1 Skipped Deployment

Project	Deployment	Preview	Comments	Updated (UTC)
langchainjs-api-refs	Ignored			Sep 25, 2025 6:02am