`anthropic-sdk-python` `v0.100.0` adds Managed Agents, webhooks

The Python SDK now exposes Managed Agents pieces including multi-agents, outcomes, webhooks, and vault validation. Useful if you want Anthropic-side agent orchestration without wiring every control path yourself, but it matters only if that stack is already on your roadmap.

• —Managed Agents support lands in the SDK with multi-agents and outcomes, so higher-level agent flows are now reachable from Python without custom wrappers.
• —webhooks support is included and the release also fixes webhook configuration, which lowers friction for async runs and event-driven backends.
• —vault validation was added alongside the agent features, suggesting Anthropic is tightening the config and secret-handling path around managed execution.
• —This is a capability unlock, not a model or pricing change. Immediate impact is moderate unless you're already building around Anthropic-hosted agents.