Architecture from requirements
Translate a player-visible outcome into data, runtime, asset, or Blueprint responsibilities before choosing tools.
A source-first course for turning one reversible idea into a tested release. Learn the lightest data and runtime paths first, then move into cooked Blueprint content and evidence-based publishing.
The course branches into runtime, data, and cooked-content paths, then brings every format back through the same release discipline.
Turn an idea into a small, maintainable architecture and choose the right format before installing a heavy toolchain.
Scaffold a UE4SS Lua mod, observe one reflected function safely, and design around lifecycle and game-thread constraints.
Make a narrow schema-backed data edit, validate it, and reduce the fields and rows that can conflict with another mod.
Establish a pinned Unreal baseline, create a LogicMod entry point, and audit the cooked output instead of shipping the project.
Test client/server and save boundaries, build a valid Info.json package, and publish evidence-based compatibility claims.
Translate a player-visible outcome into data, runtime, asset, or Blueprint responsibilities before choosing tools.
Work with reflected functions, callback frequency, object validity, the game thread, bounded retries, and idempotent reloads.
Patch only owned fields with schemas, or isolate cooked content behind a unique ModActor and nonzero chunk.
Decide what the client, host, and dedicated server own; document whether changes touch saves or leave durable content behind.
Build a matrix for clean worlds, copied saves, reload, reconnect, conflict, update, disable, uninstall, and dependency removal.
Ship the output rather than the project, validate Info.json, audit contents, version updates, and write useful rollback notes.
You do not need an Unreal project on day one. You do need a current game build, backups, version control, fresh logs, and a result you can prove in one test.