LgrappaG/Workflows-Agents

LgrappaG 资料

最近记录的来源活动: 2026年3月29日 22:06

SkillsMP 收录数据更新: 2026年10月17日 00:09

已收集 skills: 593

GitHub 星标: 2
GitHub Forks: 0

仓库 README

仓库 README 来源：README.md

查看 README 源文件

展开完整 README

# Workflows Agents — Unity Multi-Platform Toolkit v10 A comprehensive AI automation framework for **Unity game development across all platforms** (VR/XR, Mobile, PC, WebGL, Console), featuring **63 workflows + 585 AI-powered specialist skills** organized in **18 categories**, backed by a **48-agent organizational hierarchy** and **8-gate validation system** for rapid development and production-ready systems. **NEW in v10.0.0:** 🎯 **End-to-End Orchestration System** - 4,446+ LOC autonomous multi-agent orchestration with intelligent error recovery, consensus voting, and production monitoring. 29/29 tests passing (100%). See [CHANGELOG.md](CHANGELOG.md) for details. **NEW in v9.0.3:** 🎉 **Skills Enrichment Release** - 585 total skills with domain-specific YAML enrichment across 40 core skills. Measurable mandates, concrete response patterns, real usage contexts. **Previous:** v8.2.0 - Skills Expansion (450 skills), v8.1.0 (300 skills), v8.0.0 (150 skills). Inspired by and compatible with [antigravity-awesome-skills](https://github.com/sickn33/antigravity-awesome-skills). **Quick Navigation:** - 🤖 [Orchestration System](#-end-to-end-orchestration-system) - Multi-agent autonomous execution - 📖 [Quick Start Guide](docs/QUICK_START.md) - 5 minute onboarding - ⚡ [Token Optimization](docs/TOKEN_OPTIMIZATION.md) - 30-50% context savings - 🐛 [Troubleshooting](docs/TROUBLESHOOTING.md) - Common issues & fixes - 📚 [Full Documentation Index](docs/DOCUMENTATION_INDEX.md) - All references - 📊 [Version Manifest](VERSION_MANIFEST.json) - Complete inventory - 🔗 [CHANGELOG.md](CHANGELOG.md) - Full version history --- ## 🤖 End-to-End Orchestration System **Status: ✅ Production Ready** Complete multi-agent orchestration system enabling autonomous goal execution with intelligent error recovery and multi-agent consensus. ### System Overview ``` INPUT: High-level goal (e.g., "deploy-game-release") ↓ ORCHESTRATE: Auto-decompose → Resolve dependencies → Route to agents ↓ EXECUTE: Deploy, Sync, Validate, Knowledge agents coordinate ↓ MONITOR: Error detection, pattern learning, strategy optimization ↓ RECOVER: Intelligent recovery with circuit breakers, exponential backoff ↓ OUTPUT: Execution report with metrics, errors, recovery details ``` ### Key Components | Component | Purpose | Tests | |-----------|---------|-------| | **Foundation** | Orchestrator, State Manager, Heartbeat | 4/4 ✅ | | **Core Agents** | Deploy, Sync, Validation, Knowledge | 9/9 ✅ | | **Intelligence** | Error Detection, Learning, Recovery, Routing | 11/11 ✅ | | **Integration** | Multi-goal workflows, Cross-environment, E2E testing | 9/9 ✅ | | **TOTAL** | 4,446 LOC with production monitoring | **29/29 ✅** | ### Quick Commands ```bash # Initialize orchestration system python orchestration/orchestration_main.py init # Execute a goal (e.g., deploy game release) python orchestration/orchestration_main.py deploy-game-release # Check system status python orchestration/orchestration_main.py status # View available goals python orchestration/orchestration_main.py goals # Run end-to-end tests pytest tests/test_phase4_e2e.py -v ``` ### Supported Goals **Deployment Pipeline** - `deploy-game-release`: Full build → package → deploy → verify - `deploy-hotfix`: Quick hotfix deployment - `deployment-verification`: Verify successful deployment **Quality Assurance** - `validate-before-deploy`: 8-gate quality validation - `validate-all-skills`: Quality check all skills - `gate-approval-decision`: Final approval decision **Version Control** - `sync-with-remote`: Sync with git remote - `resolve-git-conflicts`: Handle merge conflicts - `sync-team-collaboration`: Multi-user sync **Learning & Optimization** - `learn-and-optimize`: Analyze metrics and suggest improvements - `analyze-metrics`: Deep metric analysis - `suggest-improvements`: Generate optimization recommendations ### Agents 1. **Deploy Agent** (250 LOC) - 4-stage pipeline: Build → Package → Deploy → Verify - Error recovery: retry_with_verbose, retry_with_backoff, rollback 2. **Sync Agent** (250 LOC) - Git operations with conflict resolution - Smart merge strategy (rebase vs merge) 3. **Validation Agent** (250 LOC) - 8-gate quality system with adaptive thresholds - Gates: Syntax, Naming, Description, Risk, Mandates, Response, Size, Consistency 4. **Knowledge Agent** (250 LOC) - Learning from execution history - Threshold adjustment recommendations - Optimization suggestions ### Intelligence Modules 1. **Error Detection Engine** (380 LOC) - Pattern recognition from error logs - Anomaly detection with statistical thresholds - Error chain identification 2. **Self-Correction Engine** (380 LOC) - Learning from successful recovery strategies - Confidence scoring for strategy selection - Adaptive planning and recommendations 3. **Recovery Manager** (500 LOC) - Circuit breaker pattern implementation - Exponential backoff with jitter - Multi-step rollback sequences 4. **Intelligence Router** (450 LOC) - Multi-agent consensus voting (4 strategies) - Conflict resolution by priority, confidence, or voting - Full audit trail of decisions ### Architecture **Environment Awareness** - Detects execution environment and adjusts behavior - Adaptive thresholds for different contexts - Optimized for strict and lenient validation modes **Error Recovery Flow** ``` Error Detected ↓ ErrorDetectionEngine (Pattern analysis) ↓ SelfCorrectionEngine (Strategy learning) ↓ RecoveryManager (Retry logic with circuit breaker) ↓ IntelligenceRouter (Multi-agent consensus) ↓ Execution (Deploy recovery or escalate) ``` **Metrics Collection** - Per-agent execution statistics - Error pattern tracking - Recovery strategy effectiveness - Performance benchmarking ### File Structure ``` orchestration/ ├── engine/ │ ├── central_orchestrator.py (main routing) │ ├── state_manager.py (persistence) │ ├── heartbeat_engine.py (monitoring) │ └── goal_decomposer.py (sub-goal decomposition) ├── agents/ │ ├── base_agent.py (abstract contract) │ ├── deploy_agent.py (deployment) │ ├── sync_agent.py (version control) │ ├── validation_agent.py (quality) │ └── knowledge_agent.py (learning) ├── intelligence/ │ ├── error_detection_engine.py │ ├── self_correction_engine.py │ ├── recovery_manager.py │ └── intelligence_router.py ├── config/ │ ├── orchestration-config.yaml │ └── agent-mapping.yaml ├── error_handlers.py (shared error types) └── orchestration_main.py (CLI) tests/ ├── test_foundation.py ├── test_agents.py ├── test_intelligence.py └── test_e2e.py ``` ### Production Features ✅ **Autonomous Execution**: Accept goals → Decompose → Execute ✅ **Error Recovery**: Detect → Learn → Recover autonomously ✅ **Consensus Voting**: Multi-agent consensus on decisions ✅ **State Persistence**: Checkpoint recovery from failures ✅ **Performance Metrics**: Comprehensive execution tracking ✅ **Adaptive Thresholds**: Adjusts for different environments ### Test Coverage ``` Foundation: 4/4 tests pass ✅ Core Agents: 9/9 tests pass ✅ Intelligence: 11/11 tests pass ✅ E2E: 9/9 tests pass ✅ ───────────────────────────────────── TOTAL: 29/29 tests pass ✅ ``` --- ## 🔒 Enterprise Hardening - Phase 1 **Status: ✅ Production Ready** Security hardening layer protecting all agent execution with policy enforcement, resource isolation, safe file operations, and circuit breaker pattern. Transparent integration — no agent code changes required. ### Security Architecture Three-layer defense-in-depth architecture: ``` LAYER 1: Policy Enforcement (CentralOrchestrator) └─ Question: "Is agent allowed to execute this goal?" └─ Decision: YES → Continue, NO → PermissionError └─ Audit: Log all policy decisions LAYER 2: Resource Isolation (AgentSandbox) └─ CPU Limits: 25-100% configurable per tier └─ Memory Limits: 256MB-1024MB configurable per tier └─ Timeout: 60-600 seconds configurable per tier LAYER 3: Safe Operations (StateManager) └─ File validation: Check path whitelist └─ Backup creation: Before every write └─ Atomic writes: Temp file + rename └─ Audit trail: Every operation logged ``` ### Security Components | Component | Purpose | Status | |-----------|---------|--------| | **SafeFileOperations** | File I/O with validation + audit trail | ✅ 340 LOC | | **SafeProcessExecution** | Subprocess execution with shell injection prevention | ✅ 340 LOC | | **CircuitBreaker** | Failure detection preventing cascading failures | ✅ 368 LOC | | **PolicyEngine** | Access control enforcement (agent/operation/path) | ✅ 316 LOC | | **AgentSandbox** | Resource limits (CPU, memory, timeout) | ✅ Configurable | ### Hard Boundaries (Cannot Bypass) ✅ **File System Protection** - Cannot write to: `/etc`, `/sys`, `/proc`, `/root`, `~/.ssh`, `~/.kube`, `~/.aws` - Can only write to: `.agents/skills`, `.agents/workflows`, `.agents/state`, `.agents/logs` ✅ **Process Execution Limits** - Max memory: 256-1024MB per tier (configurable) - Max CPU: 25-100% per tier (configurable) - Max timeout: 60-600 seconds per tier (configurable) ✅ **Agent Authorization** - Allowed agents: DeployAgent, SyncAgent, ValidationAgent, KnowledgeAgent - Each agent has specific allowed operations ✅ **Circuit Breaker Protection** - Opens after 5 failures in 60 seconds - Prevents infinite loops and cascading failures - Half-opens after 60 seconds to test recovery ✅ **Audit Trail** - Every file operation logged (path, size, backup status, result) - Every policy decision logged (allowed/denied, reason, timestamp) - Location: `.agents/logs/audit.jsonl` ### Configuration **Sandbox Resource Tiers:** ```yaml standard: 256MB memory, 25% CPU, 60s timeout premium: 512MB memory, 50% CPU, 300s timeout critical: 1024MB memory, 100% CPU, 600s timeout ``` Override in `.agents/config/hardening-policy.yaml`: ```yaml agents: DeployAgent: enabled: true allowed_operations: [deploy, validate, status, rollback] constraints: max_concurrent: 3 max_memory_mb: 512 max_cpu_percent: 50 timeout_sec: 300 ``` ### Production Features ✅ **Transparent Integration** - Existing agents work unchanged ✅ **Fallback Mode** - Direct execution if hardening disabled (via env var) ✅ **Policy Enforcement** - Pre-flight checks in CentralOrchestrator.execute_goal() ✅ **Safe Persistence** - All StateManager writes protected with SafeFileOperations ✅ **Comprehensive Audit** - Full operation trail for compliance ✅ **Defense-in-Depth** - Multiple security layers, no single point of failure ### Quick Commands ```bash # Initialize hardening system python orchestration/orchestration_main.py init # Execute with hardening enabled (default) python orchestration/orchestration_main.py deploy-game-release # Disable hardening for development (if needed) export HARDENING_ENABLED=false python orchestration/orchestration_main.py deploy-game-release # View audit trail tail -f .agents/logs/audit.jsonl # Check hardening status python orchestration/orchestration_main.py status ``` ### Integration Points Hardening is automatically applied at 3 key orchestration locations: 1. **CentralOrchestrator.execute_goal()** - Policy check before goal execution 2. **StateManager._persist_state()** - Safe file operations for all state writes 3. **StateManager.checkpoint()** - Automatic backups for recovery ### Test Coverage ``` SafeFileOperations: 5/5 tests pass ✅ CircuitBreaker: 4/4 tests pass ✅ PolicyEngine: 6/6 tests pass ✅ AgentSandbox: 2/2 tests pass ✅ Integration: 3/3 tests pass ✅ ───────────────────────────────────── TOTAL: 20/20 tests pass ✅ ``` ### Files ``` .agents/security/ ├── safe_operations.py (SafeFileOperations + SafeProcessExecution) ├── circuit_breaker.py (CircuitBreaker + CircuitBreakerManager) ├── policy_engine.py (PolicyEngine + AgentPolicy) ├── hardened_orchestration.py (Integration decorators + helpers) └── __init__.py .agents/config/ ├── hardening-policy.yaml (Agent policies + constraints) └── hardening-defaults.yaml (Resource limits, circuit breaker settings) .agents/orchestration/tests/ └── test_phase1_hardening.py (20 comprehensive tests) ``` ### Documentation **Detailed hardening documentation:** `.agents/HARDENING_PHASE1_INTEGRATION.md` --- ## 🎮 The MCP Enhancement Suite ### 🧠 Modules & Automated Usage | Module | What the Agent Can Do | Core Files | |---|---|---| | **Logic & Parkour** | Create survival stats, drain rates, and raycast-based movement. | `SurvivalHandler.cs`, `ParkourController.cs` | | **Narrative** | Lock/Unlock player input and manage Dialogue/Cinematic states. | `NarrativeDirector.cs` | | **Asset Intel** | Paint worlds with prefabs using Poisson-Disk scattering. | `WorldScaffolder.cs` | | **AI-NPC** | Give NPCs traits and goals (GOAP) that solve world state problems. | `NPCBrain.cs`, `GPlanner.cs` | | **Social Manager** | Mark "Points of Interest" and coordinate crowd behaviors. | `SocialManager.cs`, `POITag.cs` | | **Audio Intel** | Automate biome ambient sounds and material footsteps (Foley). | `AmbienceManager.cs`, `AutoFoley.cs` | | **Multiplayer** | Sync variables and trigger RPCs across the network instantly. | `NetworkSurvivalHandler.cs`, `ActionSyncer.cs` | | **UI/UX Vibe** | Theme the entire game UI and bind bars to stats without code. | `UIPalette.cs`, `UIBarBinding.cs` | | **VFX & Juice** | Add camera shake, impact feel, and mood-based lighting/fog. | `JuiceScaffolder.cs`, `AtmoController.cs` | ### 🛠️ Quick Start Command Examples - *"Add a **Hunger** stat to the player and link a **UI bar** to it."* - *"Create an **NPC** with the **Coward** trait that avoids the player."* - *"Initialize a **Dark Forest** atmosphere and set the **Ambience**."* - *"Sync the **Death Animation** for all players in multiplayer."* --- ## 🧰 Workflows (63 total) ### Core Development | Workflow | Use When | |---|---| | `/vibe-project-init` | Bootstrap a new Unity VR project from zero | | `/unity-2d-setup` | Bootstrap a Unity 2D project (URP 2D, Pixel Perfect, Sprite Atlas) | | `/unity-3d-setup` | Bootstrap a Unity 3D project (URP, NavMesh, Cinemachine 3D, LOD) | | `/unity-xr-ar-setup` | Set up a new XR/AR Unity project | | `/vibe-vr-scaffold` | Scaffold folders for a new VR feature | | `/unity-brainstorm-feature` | Facilitated design session before any feature implementation | ### Code Quality | Workflow | Use When | |---|---| | `/code-review-swarm` | Deep 4-role code review with severity report | | `/csharp-format-project` | Format all C# files | | `/unity-docs-generator` | Add XML documentation to C# scripts | | `/unity-so-architecture` | ScriptableObject event/data architecture | | `/unity-record-adr` | Write Architecture Decision Records | | `/skill-audit` | Audit all skills against the 8-gate quality checklist | | `/technical-debt-audit` | Identify and prioritize technical debt | ### Testing & QA | Workflow | Use When | |---|---| | `/vibe-vr-tdd-setup` | Scaffold EditMode/PlayMode test suite | | `/ui-testing-framework` | Set up automated UI testing | | `/load-testing-setup` | Configure load/stress testing | | `/qa-plan` | Create comprehensive QA plan | | `/i18n-testing-workflow` | Test internationalization | ### Performance & Optimization | Workflow | Use When | |---|---| | `/unity-performance-audit` | Profile CPU/GPU, fix GC allocators, verify on Quest | | `/unity-profile-audit` | AI-led performance audit using real-time MCP data | | `/unity-asset-audit` | Scan project for high-impact assets and optimization tips | | `/build-size-optimization` | Reduce build size | | `/shader-optimization-guide` | Optimize shaders for mobile/VR | | `/memory-profiler-advanced` | Deep memory analysis | | `/assets-bundle-strategy` | Configure asset bundles | | `/mobile-optimization-audit` | Audit mobile/Quest performance | ### Debugging & Recovery | Workflow | Use When | |---|---| | `/unity-smart-debug` | Diagnose Unity Console errors | | `/unity-mcp-check` | Verify direct Unity Editor connection | | `/unity-clean-project` | Clear Unity caches (Library, Temp) | | `/error-recovery` | Recover from git conflicts, broken builds | | `/project-health-check` | Overall project health assessment | ### Build & Deployment | Workflow | Use When | |---|---| | `/unity-build-quest` | Build and deploy Meta Quest APK | | `/android-specific-setup` | Android SDK, Gradle, deployment | | `/ios-specific-setup` | iOS provisioning, Xcode setup | | `/webgl-build-setup` | WebGL HTML5 deployment | | `/console-build-setup` | PS5/Xbox deployment | ### Git & CI/CD | Workflow | Use When | |---|---| | `/vibe-git-push` | Commit and push with AI-generated message | | `/vibe-git-sync` | Pull/sync from remote | | `/github-actions-unity-setup` | Set up CI/CD pipelines for Unity | | `/changelog-generator` | Generate changelog from git commits | | `/devops-audit` | Audit DevOps setup | ### Multiplayer & Backend | Workflow | Use When | |---|---| | `/multiplayer-setup` | Set up Netcode, lobbies, matchmaking | | `/backend-setup` | Set up APIs, databases, cloud services | ### Localization & Analytics | Workflow | Use When | |---|---| | `/localization-setup` | Multi-language framework setup | | `/analytics-integration` | Firebase, Amplitude, Sentry integration | ### Accessibility & Compliance | Workflow | Use When | |---|---| | `/accessibility-audit` | WCAG, VR comfort, inclusive design | | `/accessibility-audit-workflow` | Detailed accessibility review | | `/vr-accessibility` | VR comfort settings, subtitles, colorblind modes | | `/gdpr-compliance-setup` | GDPR, CCPA, COPPA compliance | ### Security | Workflow | Use When | |---|---| | `/security-audit` | Security vulnerabilities, anti-cheat | | `/security-vulnerability-scan` | Automated vulnerability scanning | ### Game Design | Workflow | Use When | |---|---| | `/game-balance-review` | Review difficulty curves, economy, retention | | `/animation-state-machine` | Set up Animator, blend trees, IK | | `/audio-setup` | Configure spatial audio, music, ambience | ### Skill & Workflow Management | Workflow | Use When | |---|---| | `/create-skill` | Create a new AI skill (with quality checklist) | | `/create-workflow` | Create a new workflow | | `/build-dev-cli` | Build a custom dev CLI tool | | `/skill-pipeline` | Define and execute skill chains | | `/project-onboarding` | Onboard new team members | | `/dependency-health-check` | Check Unity packages for updates | | `/unity-ui-scaffold` | Scaffold VR-optimized UI layouts | | `/unity-smart-placement-setup` | Configure procedural placement system | Use `/vibe-router` to describe your goal in plain English — it picks the right workflow or skill automatically. --- ## 🧠 Expert Skills (28 total) ### Unity Specialists | Skill | Use When | |---|---| | `@unity-2d-expert` | Sprite/Tilemap, 2D Physics, URP 2D, Cinemachine 2D | | `@unity-3d-expert` | URP/HDRP, NavMesh AI, 3D Physics, Cinemachine 3D, LOD | | `@unity-architect` | Architecture decisions, trade-off analysis, ADR writing | | `@unity-debugger` | Systematic GC/frame drop/judder/null-ref diagnosis | | `@unity-tdd-expert` | Writing or fixing Unity EditMode/PlayMode tests | | `@unity-smart-placement` | Procedural generation, bounds snapping, pivot fixing | ### Code & Review | Skill | Use When | |---|---| | `@csharp-master` | C# code, Unity scripting, GC/memory issues | | `@code-reviewer` | Production code review with "Ready to merge?" verdict | | `@brainstorming` | Design-first planning — validates ideas before implementation | ### VR/XR & Performance | Skill | Use When | |---|---| | `@vr-xr-specialist` | XRI v3, Quest optimization, hand tracking, spatial UI | | `@profiling-specialist` | CPU/GPU profiling, memory analysis, frame timing | | `@shader-optimizer` | Shader performance, GPU optimization, mobile/VR shaders | | `@physics-expert` | Physics optimization, collision debugging, rigidbody issues | ### Multiplayer & Backend | Skill | Use When | |---|---| | `@netcode-specialist` | Netcode for GameObjects, Photon, Mirror, state sync | | `@backend-specialist` | Server architecture, database design, API design | | `@addressables-specialist` | Asset loading, Addressables system, DLC delivery | ### Platform & Mobile | Skill | Use When | |---|---| | `@mobile-expert` | iOS/Android optimization, device fragmentation | | `@devops-engineer` | CI/CD pipelines, cloud infrastructure, deployment | ### Content & Design | Skill | Use When | |---|---| | `@game-design-specialist` | Game systems, balance mechanics, economy design | | `@animator-specialist` | Rigging, blend trees, motion capture, network sync | | `@audio-designer` | Spatial audio, mixing, FMOD integration | | `@timeline-specialist` | Cutscenes, cinematics, scripted sequences | | `@localization-expert` | Multi-language support, translations, RTL text | ### Quality & Compliance | Skill | Use When | |---|---| | `@accessibility-expert` | WCAG 2.1, colorblind modes, motor accessibility | | `@security-specialist` | OWASP Top 10, encryption, vulnerability scanning | | `@qa-strategist` | Test planning, automation frameworks, coverage | ### Meta Skills | Skill | Use When | |---|---| | `@skill-improver` | Iterative audit-fix loop for any SKILL.md | | `@workflow-assistant` | Routing, running, and creating workflows | > **Pro tip**: Start with `@brainstorming` → validate with `@unity-architect` → implement with `@csharp-master` → review with `@code-reviewer`. --- ## 💻 Install ```bash # Inject into any Unity project root git clone https://github.com/LgrappaG/Workflows-Agents.git .agents ``` Open Antigravity (or any compatible AI assistant) in the workspace — workflows and skills are auto-detected.

正在加载仓库 README…

这个仓库中的 skills

软件开发工程师特效艺术家和动画师数据科学家网页与数字界面设计师

23 个职业分类 · 已分类 100%

软件开发工程师

Implement frame-by-frame stepping for detailed animation inspection

原文语言：英语

2026-03-29

已展示 40 / 593 个已收集 Skill。

用 AI 助手安装

复制这段 Prompt，发给你正在使用的 AI 助手。

Follow https://skillsmp.com/skill-install/prompt.md to install Agent Skills from https://github.com/LgrappaG/Workflows-Agents.