Anthropic Python SDK v1.13.0 adds Managed Agents workflows, multiagent config and thread status filtering

Anthropic's Python SDK v1.13.0 adds workflows, multiagent configuration and thread status filtering to the Managed Agents API surface. The release also adds types for the Chat and Cowork unified analytics metrics API. It is an SDK update only and does not introduce a new model.

What changed in v1.13.0

The release was published by the anthropic-sdks-releaser account on the anthropics/anthropic-sdk-python GitHub repository, tagged at commit b4b7916. The release notes list two features and one chore.

Features

• Analytics types: The SDK adds types for the Chat and Cowork unified analytics metrics API.
• Managed Agents: The SDK adds workflows, multiagent configuration and thread status filtering to Managed Agents.

Chores

• Internal: The default API version header value is now named in the code. This is a refactor and is not described as a behavior change.

The release carries two assets.

What the release notes do not say

The notes are terse, and several details are not disclosed:

• Field-level schemas for the new workflow and multiagent configuration parameters are not described in the notes.
• The notes do not say which thread statuses can be filtered.
• It is not stated whether the new Managed Agents features are generally available or in beta. Developers should check the API documentation and the SDK's type definitions before relying on them.
• No breaking changes are listed.
• No model versions, context window changes, pricing changes or benchmark results are part of this release. Pricing is not applicable.

The notes show a release timestamp of 09 Oct, 15:28, without a year or time zone.

What this means

The notable change is the expansion of Managed Agents. Workflows and multiagent configuration point to Anthropic extending that surface from single-agent use toward orchestrating several agents within one managed setup. That reading is inferred from the feature names alone, as the release notes give no design details. Thread status filtering is a practical addition for anyone running many agent threads, since it lets clients query by state rather than fetching and filtering locally.

The analytics types matter for a narrower group: teams that consume usage metrics across Chat and Cowork. Typed responses reduce the need for hand-written models and make the data easier to integrate into internal dashboards.

For developers, this is a minor-version feature release. Teams already using Managed Agents should review the SDK diff and the API reference to confirm the exact parameters. Everyone else can treat it as a routine upgrade. Because the notes list no breaking changes, upgrades should be low risk, but pinning versions and running existing tests first remains standard practice.