Ghostbox — AI skill for Claude Code
Distributed AI agent orchestration - self-evolving ghosts in sandboxed containers.
How to install Ghostbox
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open kingbootoshi/ghostbox and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Ghostbox does
Distributed AI agent orchestration - self-evolving ghosts in sandboxed containers.
Alternatives in AI
- Claude Flow — An enterprise-grade AI orchestration platform that revolutionizes how developers build with AI 22.2k ★
- Open Multi Agent — TypeScript multi-agent orchestration engine — one runTeam() call from goal to result 5.8k ★
- Orchestration Skills — Multi-agent orchestration patterns 5.3k ★
README
Ghostbox
Persistent AI agents in isolated Docker containers. Each ghost gets its own vault, its own memory, and evolves across sessions.
What is this
Ghostbox lets you create isolated AI agents, each with their own sandboxed environment and persistent vault. Agents self-evolve - building tools, refining instructions, growing their knowledge base - and save their updates to GitHub.
Each agent has:
- A vault - persistent git-backed filesystem that survives restarts
- A memory system - warm facts injected into every session + deep knowledge searchable on demand
- Full coding capabilities - file I/O, bash, package installation, server hosting
- Self-evolution - agents build their own tools, refine their own instructions, grow their knowledge base
The host orchestrator manages container lifecycle, routes messages via Telegram or the REST API, and handles git persistence. A native macOS app provides a local interface.
Quick start
npx @bootoshi/ghostbox init
This runs the setup wizard: configures API keys, builds the Docker image, and creates your first ghost.
# Spawn a ghost
ghostbox spawn researcher --model anthropic/claude-sonnet-4-6
# Talk to it
ghostbox talk researcher "Explore the codebase and save what you learn"
# Start the Telegram bot
ghostbox bot
Architecture
Host (Bun) Docker Containers
+------------------------------------+
| CLI / API / Telegram | +------------------------+
| | | ghostbox-researcher |
| +-----------+ +---------------+ | | /vault (mounted) |
| | telegram |->| orchestrator |--+-->| ghost-server.ts |
| +-----------+ +---------------+ | | Pi Agent SDK |
| | api.ts |->| | | | memory + qmd |
| +-----------+ +----+-----
Related Skills
Dsh Auto Evolve
A self-evolving plugin for DeepSeek Harness (dsh). It observes how the agent runs, proposes improvements to it
Full Stack Chat Application With Multi Platform Clients AI Agent Integration
A Claude Code-style AI agent harness in Python. Uses Ollama + native LLM tool-calling to read/write files and
Archon Autonomous Workflows
Autonomous distributed AI agent platform — Gemma plans, Claude builds, self-corrects via retry loop with pgvec
Gitmachine
Git-aware sandboxed VM orchestration for AI agents
Dev Sandbox
Run untrusted code and AI agents in isolated Podman containers with krun microVMs. Network control, SSH, Privo
Phantom
An AI co-worker with its own computer. Self-evolving, persistent memory, MCP server, secure credential collect
Related Agents
Knowledge Reviewer
Reviews a diff against the engineering knowledge in .agents/knowledge/ (topics: distributed-correctness, failu
System Builder
Self-evolving platform agent. Researches APIs, generates new MCP servers, agents, and skills following existin
Infra Impl
Use for infrastructure and deployment work: containers, orchestration manifests, health probes, network policy