1---2name: game-development-orchestration3description: Use when initiating, replanning, or governing a complete game project from concept through vertical slice, production, certification, and post-launch; routes specialist game skills and prevents premature full-production commitment.4metadata:5 portable: true6 compatible_with:7 - claude-code8 - codex9---1011# Game Development Orchestration1213Coordinate a game as an evidence-gated product programme. Convert the creative premise into a playable risk-reduction sequence, then route each discipline to the narrowest specialist skill.1415## Prerequisites1617Load `world-class-engineering`, `anti-ai-slop`, `lean-game-product-development`, and the project's cultural/research governance before making irreversible product decisions.1819<!-- dual-compat-start -->20## Use When2122- Starting a mobile, PC, console, web, 2D, or 3D game project.23- Converting research or a narrative premise into a game concept and production plan.24- Defining an MVP, prototype, vertical slice, milestone plan, staffing model, or backlog.25- Auditing whether a project is ready to move from discovery to production.26- Coordinating design, engineering, art, audio, QA, release, and live operations.2728## Do Not Use When2930- A bounded implementation task already has an approved design; use the relevant engine or discipline skill.31- The request is only for generic mobile application development; use the Android, iOS, or mobile-cross family.32- The game depends on unverified living-culture, historical, legal, market, or platform claims; research and verify those claims before locking content.3334## Required Inputs3536| Input | Required | If absent |37|---|---:|---|38| Player, platform, business and cultural context | yes | Run discovery and record assumptions |39| Creative pillars and non-goals | yes | Draft and obtain owner confirmation |40| Target device classes and connectivity posture | yes for mobile | Commission a device study |41| Team, schedule, cash and rights constraints | yes | Keep estimates conditional |42| Existing prototype, repository and evidence | conditional | Inspect before planning changes |4344## Workflow45461. Load `world-class-engineering`, `anti-ai-slop`, and `ai-slop-audit`.472. For culturally grounded, historical, legal, market, or current-platform work, use the Digital Research Skills Engine and preserve its source and claim trail.483. Frame the player promise, core interaction, target platform, age/education posture, business model, safety/wellbeing boundaries, and definition of success.494. Create the concept brief, design pillars, fact/fiction/permission ledger, high-risk assumption register, evidence-class map, and prototype questions.505. Slice work into paper test, technical spike, playable prototype, vertical slice, pre-production, production, alpha, beta, release candidate, launch, and support.516. Route hypotheses, prototypes and learning gates to `lean-game-product-development`; do not use a feature list as proof of viability.527. Route mechanics to `mobile-game-design` and `gameplay-systems-architecture`; route interactive narrative, agent AI/navigation and Blender production to their specialist skills; route mathematical contracts to `game-math-and-simulation`; route technology to one approved engine after the bake-off gate.538. Route graphics implementation to `real-time-game-graphics`; route player-facing HUD, art direction, game feel and children's/educational experience to the game route in `design-system-skills`. Route lifecycle requirements and traceability to `srs-skills`; do not duplicate either engine's craft.549. If advertising is a candidate, apply the non-intrusive-ad gate before vendor or placement selection: clearly label it, require explicit choice or a predictable non-critical break, cap and pace it, preserve core play and exit/save/recovery, and define a no-ads/degraded path. Apply stricter child, privacy, store and owner rules.5510. Hold a gate at every phase. Advance only when the named evidence exists; otherwise narrow scope or run another experiment.5611. Close each milestone with the repository Delivery Definition of Done evidence pack.5758## Quality Standards5960- Prove fun, comprehension, cultural safety, technical feasibility, cost, and market access separately.61- Treat the vertical slice as production-risk evidence, not a miniature content dump.62- Maintain one decision register, one dependency map, one playable build lineage, and one prioritised backlog.63- Make every milestone answer a decision. Do not fund activity without a learning or delivery outcome.64- Preserve version-sensitive engine and store facts as dated, verified constraints.65- Keep documentation, prototype, build, device, observed-player and accountable-expert evidence separate; never promote one class into another.66- Keep learning, engagement, retention, wellbeing and commercial measures separate. A retention pass fails when regret, harmful spend, inability to stop, sleep disruption, child-safety incidents, refunds or trust materially worsen.6768## Anti-Patterns6970- Starting full production from a narrative document. Fix: prototype the highest-risk interaction first.71- Calling a greybox a vertical slice. Fix: require representative art, audio, controls, pipeline, performance, and release evidence.72- Choosing an engine by preference. Fix: run the same representative scene on target devices.73- Treating research as lore decoration. Fix: keep claim, creative adaptation, consultation, and permission ledgers.74- Planning every feature as mandatory. Fix: use must/should/could/will-not and define the smallest complete player promise.75- Hiding uncertainty inside precise schedules. Fix: use ranges, confidence, dependencies, and contingency.76- Letting discipline plans diverge. Fix: reconcile them at milestone and change-control reviews.77- Calling a platform-compliant ad non-intrusive. Fix: also test expectation, interruption, accidental-tap risk, decline/recovery, frequency, wellbeing and trust in the actual game.7879## Outputs8081| Artefact | Acceptance condition |82|---|---|83| Game concept and player-promise brief | Names player, action, emotion, platform, differentiation, exclusions |84| Game design document map | Links controlled system specifications rather than one stale monolith |85| Risk and assumption register | Has owner, test, evidence, date, impact, disposition |86| Prototype and vertical-slice plan | Each build answers explicit product, cultural, technical, and commercial questions |87| Production roadmap and backlog | Scope, dependencies, roles, estimates, gates, contingency and exit criteria are visible |88| Decision and release evidence | Another team can reproduce why the project advanced |8990## References9192- [Project lifecycle and gates](references/project-lifecycle-and-gates.md)93- [Game documentation set](references/game-documentation-set.md)94- [Kaizen game production loop](references/kaizen-game-production-loop.md)95- [Narrative and gameplay playtest loop](../game-narrative-and-interactive-story-design/references/narrative-playtest-loop.md)96- [Behaviour telemetry and tuning](../game-ai-behaviour-and-navigation/references/behaviour-telemetry-and-tuning.md)97<!-- dual-compat-end -->9899## Decision Rules100101| Evidence state | Decision |102|---|---|103| Core player action is not enjoyable or understood | Iterate or stop; do not increase content |104| Art direction works but target-device frame and memory budgets fail | Simplify pipeline and content before staffing up |105| Cultural or rights gate is unresolved | Use a fictional substitute, obtain permission, or remove the content |106| Learning claim lacks an age/construct/measure/comparator | Keep it a hypothesis; obtain educational-method approval and evidence |107| Ad interrupts play, learning, narrative, save, recovery, stopping, or safety | Remove or redesign it; platform minimums do not override the Chwezi gate |108| Vertical slice meets player, pipeline, device, cost, and governance gates | Approve bounded pre-production or production |109| Budget runway cannot cover milestone plus contingency | Reduce scope or fund the gate before committing |110111## Read Next112113- `mobile-game-design` for interaction, sessions, controls, UX, accessibility, economy, and player testing.114- `lean-game-product-development` for inception, hypotheses, prototypes, learning and gate decisions.115- `game-math-and-simulation` and `real-time-game-graphics` for mathematical and rendering contracts.116- `unity-mobile-game-development`, `godot-mobile-game-development`, or `unreal-game-development` after the engine decision.117- `apple-game-platform-delivery`, `online-multiplayer-and-game-backend`, `game-build-release-engineering`, and `game-security-anti-cheat-and-abuse` for their platform, online, release, and trust boundaries.118- `game-2d-art-animation-and-vfx-pipeline`, `game-3d-asset-pipeline`, `blender-game-asset-production`, `game-narrative-and-interactive-story-design`, `game-ai-behaviour-and-navigation`, `level-world-and-content-production`, `game-data-analytics-and-live-economy`, `game-studio-delivery-and-commercial-operations`, and `game-accessibility-localisation-and-player-safety` when those production lines are in scope.119- `game-testing-polish` and `mobile-game-release-liveops` before any ship claim.