docs(add): add middleware execution flow diagram to custom.mdx

[dudududukim]

Overview

Adds a detailed middleware execution flow diagram to the custom middleware documentation to help developers understand the complete agent execution lifecycle and all middleware hook points.

Type of change

Type: Update existing documentation

Related issues/PRs

• GitHub issue: N/A
• Feature PR: N/A
• Linear issue: N/A
• Slack thread: N/A

Checklist

• I have read the contributing guidelines
• I have tested my changes locally using docs dev
• All code examples have been tested and work correctly
• I have used root relative paths for internal links
• I have updated navigation in src/docs.json if needed

(Internal team members only / optional): Create a preview deployment as necessary using the Create Preview Branch workflow

Additional notes

Changes Made

• Added: SVG diagram (agent-middleware-execution-flow.svg) to /src/oss/images/
• Updated: Added "Detailed execution flow" section under "Execution order" in custom.mdx

Motivations

The middleware documentation explains hooks conceptually but lacks a visual representation of how they interact during agent execution. This diagram visualizes:

Complete execution lifecycle from START to completion

Invocation order of all middleware hooks (before_agent, before_model, after_model, after_agent)

Relationship between hooks, model calls, and tool execution

For detailed background, see: What is Middleware in LangChain's createAgent?

Testing Results

✅ make test passed (82 passed in 0.14s)
✅ Locally verified SVG rendering without errors using docs dev
✅ Image displays correctly at http://localhost:3000/oss/langchain/middleware/custom

✅ Follows established patterns from overview.mdx (using tag with absolute path)

✅ Alt text meets accessibility requirements

Feedback Welcome

If there are suggestions for improving the diagram content, styling, or placement, please provide feedback for a modified PR.

[sydney-runkle]

Hiya! Thanks so much for your contribution.

We're super excited about LangChain v1's middleware, and there's definitely a lot to learn in terms of expected flow w/ multiple nodes and wrap + hook style nodes.

I think maybe what could be super helpful here is a few more simple diagrams. Perhaps one that touches on the wrap-style nature of middlewares, ex

M1, M2, M3 means execution order looks like M1, M2, M3 (model) M3, M2, M1

I think another diagram that's helpful could help to explain the jump + exit conditions (loop entry node, loop exit node, etc).

Happy to chat more about what might make the most sense here.

[sydney-runkle]

Going to close this for now, but feel free to open a new PR w/ the above changes if you're interested. Thanks again for the contribution.

[dudududukim]

Oh Thanks! Do you think it's worthy drawing figures with middleware execution flow or not?
It kind of official docs for developers, so im curious about your aspective as a langchain engineer:)

[sydney-runkle]

I think worth it, we probably just want to make them more simple, worth a shot!