Skip to content

Latest commit

 

History

History
224 lines (184 loc) · 10.8 KB

File metadata and controls

224 lines (184 loc) · 10.8 KB

AlterLab GameForge -- Indie Game Development Skills for Claude Code

Project Overview

AlterLab GameForge is a collection of 34 production-grade Claude AI skills purpose-built for indie game development. The skills are organized into three categories:

  • 11 Studio Agents -- Specialized roles that emulate a full game studio team
  • 20 Workflow Skills -- Structured processes for common game dev tasks
  • 3 Engine Specialists -- Deep expertise for Godot, Unity, and Unreal Engine

Additionally, 2 genre packs (Roguelike, Narrative) provide genre-specific reference material.

Every skill is grounded in established game design theory (MDA Framework, Self-Determination Theory, Flow Theory, Bartle's Player Types) and delivers structured, actionable output.

Shared Documentation

All skills reference a common knowledge base. These docs provide the theoretical and procedural foundation:

  • @docs/collaboration-protocol.md -- How agents hand off work, share context, and avoid duplication
  • @docs/game-design-theory.md -- MDA Framework, Flow Theory, SDT, Bartle taxonomy, and design lenses
  • @docs/coordination-rules.md -- Turn order, conflict resolution, escalation paths between agents
  • @docs/agent-hierarchy.md -- Reporting structure and decision authority for all 11 agents
  • @docs/coding-standards.md -- Code style, architecture patterns, and engine-specific conventions
  • @docs/workflow-guide.md -- 11-phase development lifecycle mapping all 31 skills to project phases
  • @docs/monetization-ethics.md -- Ethical monetization framework, dark pattern identification, regulatory landscape
  • @docs/engine-comparison.md -- Structured comparison matrix for Godot, Unity, and Unreal Engine
  • @docs/mcp-integrations.md -- MCP server ecosystem for game dev, engine-specific recommendations, setup guides
  • @docs/ai-native-gamedev.md -- AI tools for game development, production-ready vs experimental assessment
  • @docs/genre-pack-spec.md -- Format specification for genre-specific enhancement packs
  • @docs/skill-quality-rubric.md -- 5-dimension quality scoring system for skill evaluation

Always check these docs before answering game development questions. They contain the canonical standards for this project.

Agent Hierarchy

The studio agents follow a clear hierarchy for decision-making:

Producer (coordination, schedule, scope)
  |
  +-- Creative Director (vision, pillars, tone)
  |     +-- Art Director (visual style, assets, UI art)
  |     +-- Audio Director (music, SFX, adaptive audio)
  |     +-- Narrative Director (story, dialogue, lore)
  |     +-- Game Designer (mechanics, systems, balance)
  |     +-- Economy Designer (currency flows, monetization, virtual economy)
  |
  +-- Technical Director (architecture, performance, tools)
        +-- QA Lead (testing, bug tracking, quality gates)
        +-- UX Designer (player experience, accessibility, usability)
        +-- Accessibility Specialist (inclusive design, EAA compliance, accommodations)

Decision authority:

  • Creative conflicts (art style vs. narrative tone) -- Creative Director decides
  • Technical conflicts (performance vs. feature scope) -- Technical Director decides
  • Scope conflicts (feature cut vs. deadline) -- Producer decides
  • Cross-domain conflicts (creative vs. technical tradeoffs) -- Producer mediates, Creative Director and Technical Director negotiate
  • Monetization conflicts (revenue vs. player trust) -- Economy Designer advises, Producer decides
  • Accessibility conflicts (accommodation vs. design intent) -- Accessibility Specialist advises, Technical Director decides implementation

Skill Routing Rules

Route user requests to the correct skill based on intent:

Agent Routing

User Intent Route To
Vision, pillars, creative direction, tone game-creative-director
Architecture, performance, tech stack game-technical-director
Schedule, scope, milestones, risk game-producer
Mechanics, systems, balance, progression game-designer
Story, dialogue, lore, worldbuilding game-narrative-director
Visual style, asset pipeline, UI art game-art-director
Music, SFX, adaptive audio, mixing game-audio-director
Testing, bugs, quality gates, regression game-qa-lead
Player experience, accessibility, usability game-ux-designer
Virtual economy, currency design, monetization game-economy-designer
Inclusive design, motor/visual/auditory/cognitive accommodations, EAA game-accessibility-specialist

Workflow Routing

User Intent Route To
Starting a new game project game-start
Ideation, brainstorming, concept exploration game-brainstorm
Market research, competitive analysis, market sizing game-market-research
Reviewing a game design document game-design-review
Reviewing game code or architecture game-code-review
Planning a development sprint game-sprint-plan
Building a rapid prototype game-prototype
Running a playtest session game-playtest
Checking game balance and tuning game-balance-check
Preparing for release or store submission game-launch
Coordinating multiple agents on a task game-team-orchestrator
Evaluating scope against timeline game-scope-check
Running a sprint/milestone/project retrospective game-retrospective
Running a post-mortem or project review game-postmortem
Generating documentation from existing code game-reverse-document
Translation, localization, internationalization (i18n) game-localization-manager
Analytics, telemetry, KPIs, data-driven design game-analytics-setup
Game jam participation (48-72 hour compressed development) game-jam-mode
CI/CD pipeline setup, automated builds, deployment game-ci-pipeline
Writing a game design document section by section game-gdd-author

Engine Specialist Routing

Detect the game engine from project files and route accordingly:

Detection Signal Route To
project.godot, .gd files, *.tscn scenes game-godot-specialist
*.unity, *.cs scripts, ProjectSettings/ game-unity-specialist
*.uproject, *.cpp/*.h with UE macros, Blueprints game-unreal-specialist

If no engine is detected, ask the user which engine they are targeting before routing.

Hooks

The following hooks run automatically during Claude Code sessions:

Hook Event Purpose
session-start SessionStart Initialize session state, load project context
session-stop Stop Save session state, write session log
pre-compact PreCompact Archive context before memory compaction
validate-commit PreToolUse Enforce commit message convention before git commits
detect-gaps PostToolUse Identify missing skill coverage in agent responses
log-agent PostToolUse Record which agents were invoked and their outputs
post-compact PostCompact Restore essential context after context compaction
subagent-track SubagentStart/SubagentStop Log subagent delegation for orchestration tracking
instructions-validate InstructionsLoaded Validate required docs exist when CLAUDE.md loads
stop-failure StopFailure Save session state and guide recovery on API errors
config-change ConfigChange Detect skill/config modifications during session

Templates

The templates/ directory contains starter templates for common game development artifacts:

  • Game Design Document (game-design-document.md)
  • Game Concept (game-concept.md)
  • Architecture Decision Record (architecture-decision-record.md)
  • Sprint Plan (sprint-plan.md)
  • Art Bible (art-bible.md)
  • Sound Bible (sound-bible.md)
  • Character Sheet (character-sheet.md)
  • Economy Model (economy-model.md)
  • Level Design Document (level-design-document.md)
  • Release Checklist (release-checklist.md)
  • Game Pillars (game-pillars.md)
  • Systems Index (systems-index.md)
  • AI Content Policy (ai-content-policy.md)
  • Post-Mortem Report (post-mortem.md)
  • Playtester Survey (playtester-survey.md)
  • Competitive Analysis (competitive-analysis.md)
  • Accessibility Audit (accessibility-audit.md)
  • Jam Concept (jam-concept.md)
  • Jam Submission Checklist (jam-submission.md)
  • CI Pipeline Configuration (ci-pipeline-config.md)

Reference templates with @templates/template-name.md in skill outputs.

Commit Convention

All commits to this repository must follow this convention:

Prefix Use
feat: New skill, hook, template, or documentation
improve: Enhancement to an existing skill or workflow
fix: Bug fix, broken reference, or typo correction
docs: README, CONTRIBUTING, or docs-only changes
chore: Scripts, CI, config, or repo maintenance

Examples:

feat: add game-economy-designer agent skill
improve: expand game-designer balance section with MDA analysis
fix: correct agent hierarchy in coordination-rules doc
docs: update quick start section in README
chore: add frontmatter validation to validate script

Quality Standards

  • Agent skills: minimum 200 lines of substantive content
  • Workflow skills: minimum 150 lines of substantive content
  • Engine specialists: minimum 200 lines with engine-specific code examples
  • No stubs or placeholder sections
  • Every skill must include concrete output templates
  • All cross-references between skills must be valid
  • All skills must include argument-hint in frontmatter
  • All skills must include allowed-tools in frontmatter
  • Agent skills should include effort level (max for directors, high for leads)
  • Agent skills with isolated context should use context: fork

Session State

Active session data is stored in production/session-state/ and session logs in production/session-logs/. Both directories are gitignored. The hooks manage this state automatically.

Project Structure

AlterLab_GameForge/
  skills/
    agents/              -- 11 studio agent skills
    workflows/           -- 20 workflow skills
    engine-specialists/  -- 3 engine-specific skills
  genre-packs/           -- Genre-specific reference material (not skills)
    roguelike/           -- Roguelike design patterns, balance, brainstorm
    narrative/           -- Narrative game patterns, pacing, brainstorm
  docs/                  -- 15 shared knowledge base documents
  hooks/                 -- 11 session lifecycle hooks
  templates/             -- 20 starter templates for game dev artifacts
  scripts/               -- Validation scripts (validate.sh)
  starters/              -- Engine-specific project configs
  production/            -- Runtime session data (gitignored)
  .github/workflows/     -- CI validation pipeline
  CLAUDE.md              -- This file
  README.md              -- Project showcase
  marketplace.json       -- Plugin manifest for Claude Code
  package.json           -- Project metadata