Claude Code 2.1.132 Update

In Claude Code 2.1.132, Managed Agents and schedule policies have been significantly reorganized.

• Managed Agents documentation was added, covering multiagent sessions, per-agent threads, streams/endpoints, events, subagent permissions, and common pitfalls.
• The Outcomes documentation explains the user.define_outcome-based graded task loop, evaluation events, deliverables, interrupts, and interaction rules.
• The Webhooks documentation covers Console registration, HMAC signature verification, payload envelope, supported events, and retry and delivery behavior.
• /schedule suggestions now default to denial, requiring all of: a named future obligation, a specific point in time, and the absence of an in-session follow-up path.
• The proactive offer guidance was removed from the schedule tool description, with related rules now handled by a separate system prompt.
• The skill limit in Managed Agents onboarding was lowered from 64 to 20, and the Prompt Suggestion Generator v2 was changed to silence suggestions that predict sensitive or risky actions.
• The security monitor now allows CronCreate, CronDelete, CronList, and RemoteTrigger, and the status line's token aggregation method has also been made clearer.