Store and resume a session transcript in private S3

Use the reference S3 SessionStore to mirror session messages and resume a conversation by session ID.

Goal

Provide controlled object-store persistence and resume for agent sessions.

Follow examples/session_stores/s3_session_store.py, install boto3 separately, and configure a private bucket/prefix client. [s3]
Pass the store through ClaudeAgentOptions(session_store=store) and retain the session ID after a successful turn. [s3]
Resume with ClaudeAgentOptions(session_store=store, resume="previous-session-id") and verify ordered history. [s3]

The new turn can refer to prior conversation, permissions are limited to the target bucket/prefix, and transcript order is preserved.

If history is incomplete, inspect object ordering and multi-writer clock skew; the example notes load latency grows linearly beyond 1,000 parts.

Python 3.10+, SDK, boto3, an isolated AWS account/bucket, and least-privilege IAM.

Grant IAM only required Put/Get/List/Delete access to the target bucket/prefix; protect transcripts as sensitive private data.

This is reference code; deletion requires explicit delete_session_via_store and an S3 lifecycle policy; never store trading credentials.

These values describe the upstream repository, not popularity or likes for an individual skill or playbook.

No metrics snapshot is available yet.

License: MIT