A new engineer needs enough authority and context to deploy safely, not a folder of copied credentials. Handover should establish organisation control, scoped access, a reproducible release path, and a way to recover when the release fails.
Separate account ownership from day-to-day permissions. The founder or organisation should retain recoverable control while engineers receive the capabilities needed for their role.
Inventory the deployment chain
List repository access, CI configuration, hosting, domain, DNS, database, storage, secrets, authentication, billing callbacks, email, and monitoring. Record owners, administrators, recovery contacts, and which changes require additional review.
Document configuration names and environment purpose without writing actual secret values into the handover guide. For Next.js, the current environment-variable reference helps explain the build and browser exposure boundary, while hosting-specific behaviour needs its own documentation.
Rehearse a fictional low-impact release
Ask the new engineer to prepare a harmless visible change in the review environment, run the relevant checks, deploy through the normal path, identify the version, and perform the agreed smoke test. Keep the exercise isolated from customer-impacting operations.
Then ask them to describe and, where appropriate, rehearse returning to the prior configuration. A rollback plan must consider database compatibility and external configuration, not just the application image or commit.
Record friction points: missing access, undocumented environment values, personal-account dependencies, unknown callback URLs, or monitoring that only the previous engineer can interpret. Resolve those gaps before declaring handover complete.
Close old access deliberately
Review former staff accounts, deploy keys, personal tokens, recovery methods, and service credentials. Removing repository membership may not remove hosting or database access. Rotate credentials where required by the organisation's policy and the actual exposure.
Keep the handover evidence useful for the next person: account inventory, setup procedure, release steps, smoke test, recovery owner, and known limitations. A long screen recording can supplement those artefacts but should not be the only navigable documentation.
The tradeoff is administrative effort versus dependence on an individual. A small SaaS can keep the process lightweight, but it still needs independent proof that another authorised person can operate the release path.
The StackFlight portfolio account describes end-to-end product and deployment work. This handover exercise is proposed guidance and does not state the internal access setup of that product.
Questions about deployment handover
Should every engineer have owner access?
No. Grant the capabilities required for the role and keep organisation recovery and ownership explicit.
Is repository access sufficient?
No. Deployment often depends on external accounts, configuration, secrets, callbacks, data access, and monitoring.
What proves handover worked?
A new authorised engineer can explain the environment, complete a safe release, verify it, and follow the recovery path. Bring unresolved dependencies to a technical handover review.