S3 to R2 Migration

Point this agent at your S3-integrated codebase and ask it to plan the move to Cloudflare R2. R2 speaks the S3 API, so most code changes are an endpoint and credential swap — the real work is the parts that don't map: ACLs, storage classes, event notifications, and the bulk copy itself.

What it does

1. Map your S3 usage: SDK clients per language, presigned-URL call sites, bucket policies/ACLs, lifecycle rules, event notifications, multipart uploads, and storage classes
2. Plan a bucket-by-bucket migration: bulk copy via Cloudflare Super Slurper or rclone, a dual-write window, then cutover with a rollback path
3. Scaffold the code changes: per-SDK endpoint override to <account_id>.r2.cloudflarestorage.com, wrangler r2 bucket create commands, CORS translation, and an event-notifications → Queues/Workers sketch
4. Sandbox — round-trip put/get/list/multipart/presign operations against the R2 test bucket and report PASS/PARTIAL/FAIL with per-operation evidence

Prerequisites

• CLOUDFLARE_API_TOKEN — token with R2 edit permission on a test account
• CLOUDFLARE_ACCOUNT_ID — the account that owns the R2 test bucket
• R2_ACCESS_KEY_ID / R2_SECRET_ACCESS_KEY — S3-compatible credentials for the R2 test bucket
• AWS_READONLY_ACCESS_KEY_ID / AWS_READONLY_SECRET_ACCESS_KEY — optional; read-only inventory of the source S3 buckets

Current scope

This solution maps code, plans the migration, scaffolds the changes, and validates the R2 test bucket. It does not execute the bulk data copy itself (it generates the Super Slurper checklist or rclone commands for you to run), does not migrate Glacier archives (they must be restored first), and does not build the replacement for S3 Select or requester-pays workloads — those are flagged as redesigns.