convez.ai
SkillAutomaçãoAvançado
frankxai

Agentic Creator OS

frankxai/agentic-creator-os

SO para criadores alimentado por IA com 90+ skills auto-ativando

Ver no GitHub
BomPluginScore · de 0 a 100

Sistema operacional production-grade para assistentes de IA com 90+ skills, 65+ comandos e 38 agentes especializados que se auto-ativam conforme o contexto. Integra-se nativamente com Claude Code, Cursor, Windsurf e Gemini, oferecendo workflows reutilizáveis, hooks de segurança e roteamento inteligente de intenção.

Instalação

$npx convez-ai frankxai/agentic-creator-os

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
12
Forks
6
Acompanhando
1
Saúde
85%
Último commit
há 3 dias
Versão
v8.0-assets
Issues abertas
23
Contribuidores
6

Casos de uso

  • Automação de workflows de conteúdo (artigos, posts, SEO)
  • Desenvolvimento full-stack com agentes especializados
  • Roteamento inteligente de tarefas para múltiplos assistentes de IA

Veredito

Solução enterprise para quem precisa orquestrar múltiplos agentes de IA com segurança e contexto. Maduro tecnicamente, mas exige curva de aprendizado considerável e infraestrutura específica para máximo aproveitamento.

README

A tradução para o português deste README ainda está na fila.

Agentic Creator OS

The Operating System for AI-Powered Creators

One install. Any coding agent. 90+ skills, 65+ commands, 38 agents — auto-activating.

Agentic Creator OS — FRANK-Ω Command Center

Version License Platforms


What Is the Agentic Creator OS?

A production-grade skill and agent system for AI coding assistants. It ships skills (domain knowledge), commands (reusable workflows), agents (specialized personas), and safety hooks — configured to auto-activate based on what you're working on.

Works with Claude Code, Grok Build (xAI CLI/TUI — full native harness + .grok/skills + .grok/hooks + subagents + MCP + grok-harness-adapter + Claude compat), Cursor, Windsurf, Gemini Code Assist, Antigravity, and any AI coding agent that reads markdown context files. Built on SIP (Starlight Intelligence Protocol).

You: "write a blog post about AI agents"
  → Agentic Creator OS detects: creation + blog
  → Auto-loads: content-strategy skill
  → Routes to: /article-creator workflow
  → Result: Guided article with SEO, images, social distribution

System Architecture

Agentic Creator OS — Layered Architecture

The Agentic Creator OS is structured in four layers, each auto-activating based on context:

flowchart LR
  Intent["User intent"] --> Router["/acos smart router"]
  Router --> Skills["Auto-activating skills"]
  Router --> Commands["Slash workflows"]
  Router --> Agents["Specialized agents"]
  Skills --> Hooks["Safety hooks"]
  Commands --> Hooks
  Agents --> Hooks
  Hooks --> Output["Audited creator output"]
  Output --> Learning["Trajectory learning"]
  Learning --> Router
LayerWhat It DoesCount
SkillsDomain knowledge modules that load automatically75+
CommandsReusable workflows triggered via slash commands35+
AgentsSpecialized personas with distinct expertise38
Safety HooksCircuit breaker, audit trail, IAM, self-modify gate5

Architecture & Harness Role (Whole Stack)

ACOS (Agentic Creator OS) is the shared execution layer (skills + commands + agents + workflows + safety) used across the FrankX / Starlight ecosystem. It is harness-agnostic but provides first-class native adapters for the full fleet.

LayerRoleKey
SIP (Starlight Intelligence Protocol)Cross-repo memory, MCP substrate (github/fs-starlight/git), sovereignty, attestation, vaultsstarlightintelligence.org/protocol
ACOS90+ skills, 65+ commands, 38 agents, workflows, v10+ safety (IAM, circuit-breaker, audit, self-modify), auto-activation via skill-rules + excellenceThis repo; junctions to ~/.claude + .grok compat
Grok Build (xAI)TUI speed, subagents (explore/plan), MCP, image/video, .grok/skills + hooks native + full .claude compat. Personal excellence seeds (4 + 2 json) via grok-harness-adapteradapters/grok/ + install --platform=grok; .grok-only grok-personal layer only
Claude CodeCanonical full surface (slash commands, hooks, IAM, skill-rules)Direct .claude/ install
Other (Cursor, Windsurf, Gemini, AGY)Context files + embedded skills/agents; delegation via multi-orchestrator.cursorrules etc.
FrankX / ArcaneaBrand voice, personal creative, specific verticals (content, music, oracle, etc.)Overlays + personal .grok/.claude partitions

Core vs Grok-Personal vs Personal-Creative Partition (strict): See SHARING.md + SIP §5. The 4 excellence seeds live ONLY in .grok/ for Grok users ("a bit magical"). Everything else is core/shared via .claude/ junctions. Use sip-share-audit before cross moves. 5-fleet parity + God 99 gates (repo-mastery + gstack + santa/verification) on all work.

This separation + adapters make ACOS the universal substrate: one catalog, many harnesses, governed outputs.


Quick Start

Claude Code (Full Feature Set)

git clone https://github.com/frankxai/agentic-creator-os.git
cd agentic-creator-os
./install.sh

# Open any project with Claude Code
claude
/acos

Cursor / Windsurf

git clone https://github.com/frankxai/agentic-creator-os.git
cd agentic-creator-os
./install.sh --platform=cursor    # or --platform=windsurf

This generates a .cursorrules or .windsurfrules file with all skills embedded.

Gemini Code Assist

git clone https://github.com/frankxai/agentic-creator-os.git
cd agentic-creator-os
./install.sh --platform=gemini

This generates a GEMINI.md context file for Gemini to read on session start.

Grok Build (xAI CLI/TUI) — Full Native Harness Support

git clone https://github.com/frankxai/agentic-creator-os.git
cd agentic-creator-os
./install.sh --platform=grok
  • Generates GROK.md (ACOS briefing + Frank DNA + excellence gates) + AGENTS.md pointer
  • Seeds .grok/skills/ (exact 4 grok-personal excellence seeds per SHARING.md + SIP §5: harness-integration, excellence-review, repo-mastery, multi-harness-orchestrator) + 2 excellence json hooks as .grok/hooks/ ONLY + .grok/agents/ (core ACOS/gstack/santa/verification via .claude/ junctions + compat)
  • Grok Build auto-loads project .grok/ (highest prio) + ~/.grok/ + full ~/.claude/ compat (skills, commands, agents, hooks)
  • Native Grok features leveraged: /skills, subagents (/task with personas + worktree + resume), MCP (github + fs-starlight + git), image/video gen, TUI speed
  • Run grok in project → /hooks-trust → /skills grok-harness (or harness-integration / repo-mastery) or natural language ACOS tasks
  • Built-in excellence gates (God 99): repo-mastery (rules first), gstack (qa/browse/design-review), santa-method + verification-loop, plan-* reviews, cso
  • Multi-harness: multi-harness-orchestrator auto-detects + emits exact delegation (e.g. claude -p "..." with injected rules + DNA + partition filter)

This is the richest non-Claude experience — native skills/agents/hooks/subagents/MCP + full ACOS excellence substrate + 5-fleet parity. See adapters/grok/index.ts and .claude/skills/grok-harness/.

Any AI Coding Agent

git clone https://github.com/frankxai/agentic-creator-os.git
cd agentic-creator-os
./install.sh --platform=generic

Generates a CONTEXT.md file. Point your agent at it as a system prompt or project instructions file.

See QUICKSTART.md for detailed setup per platform


Core Concepts

The Agentic Creator OS is built around six foundational concepts. Understanding these unlocks the full system.

Skills — Auto-Activating Domain Knowledge

Skills are markdown files containing deep domain expertise. They activate automatically when the Agentic Creator OS detects relevant context in your work.

You start writing a blog post
  → skill-rules.json matches: "blog" + "write"
  → Auto-loads: content-strategy + seo-content-writer + schema-markup
  → Your AI agent now has SEO expertise, content structure, and metadata knowledge
Skill CategoryExamplesTrigger Patterns
TechnicalTDD, React patterns, TypeScript, debuggingtest, component, error
CreativeBrand voice, music mastery, book writingblog, music, create
InfrastructureOracle Cloud, Kubernetes, Terraformdeploy, cloud, infra
BusinessProduct management, financial modelingproduct, strategy, pricing

Commands — Reusable Slash Workflows

Commands are multi-step workflows triggered by a single slash command. Each command orchestrates skills, agents, and tools in sequence.

CommandWhat It DoesSteps
/studioMultimodal productionBrief → Route model → Craft → Generate (async) → Assemble
/article-creatorFull blog pipelineResearch → Outline → Draft → SEO → Images → Social
/factoryMass content productionBatch input → Parallel agents → Quality gate → Export
/specFeature specificationRequirements → Design → Task breakdown → Validation
/infogeniusVisual intelligenceTopic → Research → 4K image generation → Metadata
/create-musicMusic productionGenre → Prompt engineering → Generation → Metadata
/councilStrategic decisionsMulti-agent deliberation → Weighted synthesis → Verdict

Agents — Specialized AI Personas

Each agent has distinct expertise, voice, and tool access. The Agentic Creator OS routes to the best agent based on task type.

AgentSpecialtyWhen It Activates
Developmental EditorStory structure, pacing, narrative architectureBook/long-form writing
Music ProducerSuno prompts, genre production, audio engineeringMusic creation
Technical ArchitectSystem design, Oracle Cloud, enterprise patternsArchitecture tasks
UI/UX ExpertDesign systems, accessibility, Figma integrationFrontend work
SEO IntelligenceKeywords, citations, structured dataContent optimization
Starlight OrchestratorMulti-agent coordination, strategic decisionsComplex workflows

Smart Router — Intent-Based Routing

The /acos command analyzes your request and routes to the optimal subsystem:

/acos "create an infographic about machine learning"
  ├── Intent detected: visual + educational
  ├── Agent selected: Visual Creator
  ├── Skills loaded: infogenius, design-excellence
  ├── Command chain: /research → /infogenius
  └── Output: 4K infographic with research-grounded content

Multimodal Studio — Image · Video · Character

One connector, every modality. ACOS generates images, video, and consistent characters across 30+ frontier models through a single MCP — model-agnostic, brand-locked, and async by default.

/studio "hero image + 3 social cards + 5s teaser for the launch post"
  ├── Routes each shot to the right model (Soul/Flux → stills, Kling/Veo → motion)
  ├── Holds one character across the whole set via create_character → reuse ID
  ├── Submits assets in parallel, polls for results, downloads to canonical paths
  └── Output: a coherent, brand-locked campaign — not disconnected one-off images
ModalityModels (via Higgsfield MCP)Differentiator
ImageSoul (4K), Flux, Seedream, Nano BananaBrand tokens injected per asset
VideoKling, Hailuo, Veo, Sora-class, DoPImage→video first (cheaper, keeps composition)
CharacterSoul character trainingOne reference, identical across an entire series

Skill: multimodal-studio · Agent: Multimodal Director · Commands: /studio, /generate-video. Connect: claude mcp add --transport http --scope user higgsfield https://mcp.higgsfield.ai/mcp. Stays vendor-agnostic — any MCP filling ~~image generation / ~~video generation works (CONNECTORS.md). Full guide + the ACOS-vs-Antigravity positioning: docs/multimodal-studio.md.

Safety Hooks — Autonomous Protection

Five systems that run automatically to keep the Agentic Creator OS reliable:

HookWhat It Prevents
Circuit BreakerStops the agent from brute-forcing broken approaches (3→warn, 5→restrict, 8→block)
Audit TrailAppend-only JSONL log of every operation — tamper-proof ground truth
Agent IAMRole-based scoping — a content writer can't modify build configs
Self-Modify GateSnapshots state before config changes, auto-reverts if intelligence drops
Quality GateBlocks edits to secrets files, warns on production file modifications

Self-Learning — Gets Smarter Every Session

The Agentic Creator OS tracks tool sequences and outcomes across sessions, then injects successful patterns as context for similar future tasks. No external API — pure local intelligence.


v10 — Autonomous Intelligence

v10 introduces 5 safety-first autonomous systems that make your AI coding agent smarter over time:

1. Experience Replay

The agent remembers what worked. Past successful tool sequences are injected as context for similar new tasks.

2. Agent IAM (Identity & Access Management)

Per-profile scoping — a content writer can't modify your build config, a music producer can't run shell commands. 6 profiles with tool and directory restrictions.

3. Confidence Circuit Breaker

Tracks failures per file. After 3 failures: warn. After 5: restrict. After 8: block. Prevents the agent from brute-forcing broken approaches.

4. Conservative Self-Modify Gate

Before changing its own config, ACOS snapshots the current state and measures intelligence score. If the score drops more than 5 points after a change: auto-revert.

5. Immutable Audit Trail

Append-only JSONL logging of every tool use, IAM decision, gate verdict, and config change. The ground truth for what happened in a session.

                        /acos (Smart Router)
                            |
              +-------------+-------------+
              |             |             |
         Commands (35+)  Skills (75+)  Agents (38)
              |             |             |
              +------+------+------+------+
                     |             |
              +------+------+  +--+--+
              |             |  |     |
          Hooks (v10)   Workflows  MCP Servers
              |
    +---------+---------+---------+
    |         |         |         |
 Circuit   Audit    Self-Mod   Agent
 Breaker   Trail    Gate       IAM

What's Included

Commands (35+)

CategoryCountExamples
Creation12/studio, /generate-video, /article-creator, /create-music, /infogenius, /factory
Strategy6/starlight-architect, /council, /research, /plan-week
Development4/spec, /nextjs-deploy, /ux-design, /automation-dev
System5/acos, /inventory-status, /mcp-status, /publish
Quality3/review-content, /classify-content, /polish-content

Commands are Claude Code slash commands. On other platforms, describe the task and ACOS skills guide the agent to the same workflows.

Skills (75+ Auto-Activating)

Skills are domain knowledge modules that activate automatically based on what you're doing. No manual invocation needed.

CategoryCountExamples
Technical25+TDD, systematic-debugging, mcp-architecture, react-nextjs-patterns
Creative10+multimodal-studio, frankx-brand, suno-ai-mastery, excellence-book-writing
Content15+content-strategy, social-media-strategy, video-production
Oracle/Cloud5+oci-services-expert, oracle-database-expert, oracle-ai-architect
Business5+product-management, coaching-program, financial-modeling
System15+swarm-orchestration, agentic-jujutsu, planning-with-files
Personal5+spartan-warrior, gym-training, health-nutrition

How Auto-Activation Works:

skill-rules.json → 22 pattern rules
  "tests" + "component" → loads test-driven-development + react-patterns
  "blog" + "write"      → loads content-strategy + seo-content-writer
  "deploy" + "vercel"   → loads vercel-deployment + nextjs-best-practices

Agents (38 Specialized)

DomainAgentsExamples
Writing8Developmental Editor, Line Editor, Content Polisher, Story Architect
Strategy5Starlight Orchestrator, Visionary, Oracle AI Architect
Design4UI/UX Expert, Website Builder, Frankx Content Creation
Production5Multimodal Director, Music Production, Sonic Engineer, Nano Banana Image Gen
Business4Product Launch, Coaching Program, Product Development
Technical5Technical Architect, Oracle Database, OpenAI AgentKit
Publishing6Master Story Architect, Character Psychologist, Sensitivity Reader
System3Skill Expert, Project Discovery, Rapid Content

v10 Safety Hooks

HookTriggerPurpose
Circuit BreakerPostToolUse (failure)Track failures per file, escalate at thresholds
Audit TrailAll tool useAppend-only JSONL logging
Self-Modify GateConfig changesSnapshot + validate intelligence score
Agent IAMPreToolUseEnforce per-profile tool/path scopes
Quality GatePreToolUse (Edit/Write)Block edits to secrets, warn on production files

How It Works

The /acos Smart Router

Agentic Creator OS — Smart Router

Everything starts with /acos. It parses intent and routes to the best subsystem:

/acos "write a blog post about AI agents"
  → Detects: creation + blog
  → Routes to: /article-creator
  → Auto-loads: content-strategy skill
  → Result: Guided article creation workflow

For multi-domain requests, it chains commands:

/acos "Create a blog post about AI music with images and social posts"
  → /research "AI music production"
  → /article-creator (with research context)
  → /infogenius (hero image)
  → /generate-social (distribution)

Orchestration Patterns

PatternDescriptionBest For
PipelineSequential: Research → Plan → Create → Optimize → PublishBlog posts, newsletters
ParallelConcurrent agents with synthesisCode review, multi-analysis
Weighted SynthesisExpert voting with percentage weightsStrategic decisions

Self-Learning

Agentic Creator OS — Self-Learning Cycle

The Agentic Creator OS learns from every session through a four-stage cycle:

StageWhat Happens
1. Session StartCreates a trajectory record for the current session
2. Track OperationsMonitors tool sequences, file changes, and outcomes
3. Extract PatternsIdentifies successful workflows and tool combinations
4. Apply IntelligenceInjects learned patterns as context for future sessions
Session Start → Create trajectory record
Session Active → Track operations and outcomes
Session End → Extract patterns → Store learnings
Future Sessions → Apply accumulated intelligence

Multi-Platform Install

The install script detects your platform and configures accordingly:

./install.sh                      # Auto-detect platforms
./install.sh --platform=claude    # Claude Code only
./install.sh --platform=cursor    # Cursor only
./install.sh --platform=windsurf  # Windsurf only
./install.sh --platform=gemini    # Gemini Code Assist only
./install.sh --platform=grok      # Grok full (GROK.md + .grok/ seeds from grok-harness-adapter)
./install.sh --platform=generic   # Any agent (CONTEXT.md)
./install.sh --platform=all       # All detected platforms

What Gets Installed Per Platform

ComponentClaude CodeGrok Build (xAI)CursorWindsurfGeminiGeneric
Skills (knowledge)~/.claude/skills/.grok/skills/ + GROK.md + ~/.claude compat (via grok-harness-adapter seeds).cursorrules.windsurfrulesGEMINI.mdCONTEXT.md
Commands (workflows)~/.claude/commands/Context + native /skills + subagents————
Agents (personas)~/.claude/agents/.grok/agents/ + subagents.cursorrules.windsurfrulesGEMINI.mdCONTEXT.md
Hooks (safety)settings.json + acos/hooks.grok/hooks/ (trust with /hooks-trust) + excellence gates————
Auto-activationskill-rules.jsonharness-integration + excellence-review seeds (gstack/santa/verification)————
MCP serversOptional————

Claude Code gets the full feature set including slash commands, lifecycle hooks, and auto-activation rules. Other platforms get skills and agent definitions embedded in their context files.


Directory Structure

agentic-creator-os/
├── .claude/
│   ├── commands/           # 35+ slash commands
│   ├── skills/             # 75+ auto-activating skills
│   ├── agents/             # 38 specialized agents
│   ├── hooks/              # v10 safety hooks
│   │   ├── circuit-breaker.sh
│   │   ├── audit-trail.sh
│   │   ├── self-modify-gate.sh
│   │   └── quality-gate.sh
│   ├── agent-iam.json      # Role-based access control
│   ├── skill-rules.json    # 22 auto-activation rules
│   └── hooks.json          # Hook lifecycle config
│
├── adapters/               # Platform adapters (TS context + seed generators)
│   ├── grok/               # grok-harness-adapter: index.ts (isGrok, parse, generateGrokContext, getGrokSeeds, installGrokPlatform), GROK excellence seeds
│   ├── opencode/           # OpenCode integration adapter
│   ├── cursor/             # .cursorrules generator (planned)
│   ├── windsurf/           # .windsurfrules generator (planned)
│   ├── gemini/             # GEMINI.md generator (planned)
│   └── generic/            # CONTEXT.md generator (planned)
│
├── departments/            # Agent team definitions
│   ├── content/
│   ├── design/
│   ├── dev/
│   ├── marketing/
│   └── business/
│
├── mcp-servers/            # MCP integrations (optional)
├── templates/              # Reusable templates
├── workflows/              # Workflow definitions
├── docs/infographics/      # Visual documentation
├── install.sh              # Multi-platform installer
├── CLAUDE.md               # Claude Code context
└── package.json            # v10.1.0

The Starlight Connection

ACOS is the Claude Code implementation of the Starlight Intelligence System — a universal, platform-agnostic framework for multi-agent orchestration.

Starlight Intelligence System (Framework)
├── 5-layer cognitive architecture
├── 7 council agents with emergent leadership
├── 6 orchestration patterns
└── Platform adapters
     └── ACOS (Claude Code + Grok full harness via grok-harness-adapter + Multi-Platform)
         ├── 35+ commands routed through /acos
         └── Grok: adapters/grok/index.ts + install.sh --platform=grok (GROK.md + .grok/ seeds: excellence, repo-mastery, gstack gates)
         ├── 38 agents aligned to Starlight council
         ├── 75+ auto-activating skills
         ├── v10 safety systems (IAM, circuit breaker, audit)
         └── Self-learning via trajectory patterns

Design Principles

  1. Configuration over Code — Markdown + JSON + AI's native abilities. Zero install friction.
  2. Commands over Prompts — Reusable workflows beat one-off prompts. Build once, use forever.
  3. Agents over Chat — Specialized expertise beats general conversation.
  4. Files over Memory — Persistent artifacts beat ephemeral threads. Everything is git-trackable.
  5. Progressive Disclosure — Load only what you need. ~100 tokens metadata before 5K skill definition.
  6. Open and Forkable — MIT licensed. Every component replaceable.

Plugin Marketplace

Extend ACOS with installable domain plugins from agentic-creator-skills:

PluginPurpose
coreSession management, workspace memory, task tracking
content-engineContent creation, voice enforcement, publishing
visual-studioStyle-governed image generation, 7-gate quality filter
music-labAI music production, Suno prompt engineering
brand-architectBrand voice system, tone adaptation, terminology
product-launcherProduct specs, sprint management, deployment
intelligenceIntelligence scoring, pattern analysis, reporting
design-excellenceDesign system governance, 5-gate QA, Figma integration
# Install plugins (Claude Code)
claude plugin marketplace add frankxai/agentic-creator-skills
claude plugin install core content-engine design-excellence

ACOS = the operating system. Plugins = the app store. Install only what you need.


Version History

VersionDateHighlights
v11.0Feb 2026Plugin ecosystem, design swarm, v11 skills absorbed from acos-intelligence-system
v10.1Feb 2026Deep clean — removed claude-flow artifacts, honest metrics, clean repo
v10.0Feb 2026Autonomous Intelligence — 5 safety systems, multi-platform, Agent IAM
v8.0Feb 2026Starlight integration, GitHub release assets, swarm orchestration
v7.0Feb 2026Visionary agent, premium infographics, expanded commands
v6.0Jan 2026Smart router (/acos), auto-activation, hooks
v5.0Jan 20267-pillar architecture, 62 skills, swarm topologies
v4.0Jan 2026Initial release, 14 commands, basic skills

Credits & Attribution

ACOS builds on work from the open-source community:

ProjectAuthorWhat We Learned
claude-flow@ruvnetSwarm orchestration, hierarchical topologies
wshobson/agents@wshobsonPlugin architecture, modular skills
obra/superpowers@obraProgressive disclosure, token-efficient loading
diet103/claude-code-infrastructure-showcase@diet103skill-rules.json auto-activation pattern
ChrisWiles/claude-code-showcase@ChrisWilesHook automation patterns
Pimzino/claude-code-spec-workflow@PimzinoSpec-driven development
github/github-mcp-serverGitHubOfficial GitHub MCP server
agentic-jujutsu@ruvnetSelf-learning trajectory patterns

Built on: Claude Code by Anthropic, Model Context Protocol


Related Projects


License

MIT — Use it, fork it, build your own OS with it.


Agentic Creator OS v11.0 — The Operating System for AI-Powered Creators

90+ Skills | 65+ Commands | 38 Agents | 8 Plugins | Multi-Platform | Self-Learning

Built by FrankX — AI Architect & Creator

GitHub | Website | Issues

Ver o original no GitHub

Bom para

Relacionados