Source: github.com/langchain-ai/langgraph/releases/tag/1.2.20
The langchain-ai/langgraph repository published the 1.2.20 release, listing the changes made since 1.2.19. The headline addition is a strict resume mode, which refuses interrupt answers that are no longer current.
Another feature change gives each interrupt() call within a task its own identifier. The release also updates the error message for multiple interrupts so it links to the current documentation.
The remaining entries are dependency updates under dev dependencies, including types-requests and a major-version bump of pycryptodome for the langgraph package.
Why it matters if you build agents
Workflows that pause for human input can now reject answers given against an outdated interrupt when strict resume mode is used. Code that tracks interrupts can rely on a separate ID for each interrupt() call in a task. Check the changelog for any other behavior changes before upgrading a pinned version.
Get stories like this every weekday
One short email with the day's agent tooling news.