[FEATURE]Durable cron jobs should survive session close, not just app restart

Preflight Checklist

• [x] I have searched existing requests and this feature hasn't been requested yet
• [x] This is a single feature request (not multiple features)

Problem Statement

CronCreate durable: true jobs should survive full session close, not just application restarts

Proposed Solution

The infrastructure is already partially there � durable: true already writes to .claude/scheduled_tasks.json. The missing piece is re-hydration on session start.

On Claude Code startup (any new conversation), the runtime should:

1. Read .claude/scheduled_tasks.json
2. For each durable task, check if the scheduled time has passed:

• Passed + one-shot: fire the prompt immediately, then delete the entry
• Passed + recurring: fire immediately, then reschedule the next interval
• Not yet due: re-register in the in-memory scheduler for the upcoming fire time

1. For tasks that fired while Claude was closed, inject the stored prompt into the new conversation as if it just fired

The stored prompt already contains all the context needed � no new data model changes required. It's purely a startup loader that reads the file and re-populates the in-memory scheduler (and fires any
missed jobs) before the first user message is processed.

Alternative Solutions

_No response_

Priority

Medium - Would be very helpful

Feature Category

CLI commands and flags

Use Case Example

_No response_

Additional Context

_No response_