[PR #4039] docs: add LangGraph final-state receipt guidance #4045

Open
opened 2026-06-05 18:53:59 -04:00 by yindo · 0 comments
Owner

📋 Pull Request Information

Original PR: https://github.com/langchain-ai/docs/pull/4039
Author: @aDragon0707
Created: 5/18/2026
Status: 🔄 Open

Base: mainHead: docs/langgraph-final-state-receipts


📝 Commits (1)

  • 334911f docs: add final-state receipt guidance for LangGraph

📊 Changes

1 file changed (+30 additions, -0 deletions)

View changed files

📝 src/oss/langgraph/durable-execution.mdx (+30 -0)

📄 Description

Summary

Adds a small section to the LangGraph durable execution docs about auditing final-state claims from production agents.

The section keeps the receipt pattern boring and checkable by separating:

  • claim
  • evidence
  • authority for the next action

It also points evidence to LangGraph-native durable execution artifacts such as thread_id, checkpoint_id, task results, and interrupt approval state.

Motivation

This follows the direction discussed in langchain-ai/langgraph#7844: final messages such as Done. All tests passed. Ready to publish. can combine completion, test, and external-action claims. The docs should encourage users to attach durable evidence and make the next owner explicit instead of trusting the final assistant message itself.

Verification

  • git diff --check
  • Ran scripts/check_cross_refs.py with PYTHONPATH using the bundled Python runtime; it reported an existing unrelated unresolved cross-reference in oss/javascript/integrations/providers/openai.mdx and then hit a Windows GBK encoding issue while printing the failure marker.
  • Ran npx markdownlint-cli2 src/oss/langgraph/durable-execution.mdx; it reported pre-existing lint issues in this file, with no new issue in the added section.

🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.

## 📋 Pull Request Information **Original PR:** https://github.com/langchain-ai/docs/pull/4039 **Author:** [@aDragon0707](https://github.com/aDragon0707) **Created:** 5/18/2026 **Status:** 🔄 Open **Base:** `main` ← **Head:** `docs/langgraph-final-state-receipts` --- ### 📝 Commits (1) - [`334911f`](https://github.com/langchain-ai/docs/commit/334911fe34032e28a0327b0c0f3dfecfaf25b52b) docs: add final-state receipt guidance for LangGraph ### 📊 Changes **1 file changed** (+30 additions, -0 deletions) <details> <summary>View changed files</summary> 📝 `src/oss/langgraph/durable-execution.mdx` (+30 -0) </details> ### 📄 Description ## Summary Adds a small section to the LangGraph durable execution docs about auditing final-state claims from production agents. The section keeps the receipt pattern boring and checkable by separating: - claim - evidence - authority for the next action It also points evidence to LangGraph-native durable execution artifacts such as `thread_id`, `checkpoint_id`, task results, and interrupt approval state. ## Motivation This follows the direction discussed in langchain-ai/langgraph#7844: final messages such as `Done. All tests passed. Ready to publish.` can combine completion, test, and external-action claims. The docs should encourage users to attach durable evidence and make the next owner explicit instead of trusting the final assistant message itself. ## Verification - `git diff --check` - Ran `scripts/check_cross_refs.py` with `PYTHONPATH` using the bundled Python runtime; it reported an existing unrelated unresolved cross-reference in `oss/javascript/integrations/providers/openai.mdx` and then hit a Windows GBK encoding issue while printing the failure marker. - Ran `npx markdownlint-cli2 src/oss/langgraph/durable-execution.mdx`; it reported pre-existing lint issues in this file, with no new issue in the added section. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
yindo added the pull-request label 2026-06-05 18:53:59 -04:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langchain-ai/docs#4045