Sno Station
Sno Station은 AI 에이전트를 하나의 팀으로 묶는 오픈소스 워크스테이션으로, 로컬 머신에서 에이전트 협업을 가능하게 함
Claude Code용 다중 에이전트 거래·스웜 인텔리전스·GitHub 자동화 플러그인, 68개 에이전트 기반 19개 프로덕션 플러그인
Multi-agent trading, swarm intelligence, and GitHub automation plugins for Claude Code. 19 production-grade plugins built from 68+ specialized agents.
플러그인Python · ShellGitHub @jmanhype직접 설치운영자 미확인
써 보셨어요? 한 번만 눌러 주세요. 로그인은 필요 없어요.
개인 계정의 공개 저장소 소유자는 GitHub로 로그인해 운영자를 확인하고 소개를 직접 관리할 수 있습니다. 관리자 승인은 필요하지 않습니다.
로그인한 뒤 확인 파일을 만들어 이 저장소의 기본 브랜치에 30분 안에 커밋해 주세요. 다른 브랜치나 포크의 파일은 인정하지 않습니다. 파일을 올린 뒤 아래에서 확인하면 관리자 승인 없이 관리할 수 있습니다. 자격을 유지하려면 파일을 남겨 두세요.
로그인이 필요하면 로그인 화면으로 이동합니다. 로그인 후 이곳으로 돌아와 원하는 확인 버튼을 다시 눌러 주세요.
설치형 프로젝트는 배포 도메인 없이도 확인할 수 있습니다. 개인 저장소는 GitHub 로그인으로, 조직 저장소는 위의 저장소 확인 파일로 직접 확인해 주세요.
README 전체를 불러오는 중입니다. 아래는 저장된 발췌입니다.
Collection of 23 plugins for Claude Code. Each plugin is a directory under plugins/ containing a plugin.json manifest, a README.md, and in some cases implementation code. Mixed. Some plugins (quant-trading-system, code-safety-monitor, ace-context-engineering) contain implementation code. Most plugins are manifest + README definitions that describe agent prompts and workflows. No centralized test suite. [README 설치·사용 부분 발췌] InstallationREADME 전문 보기 ↗
Collection of 23 plugins for Claude Code. Each plugin is a directory under plugins/ containing a plugin.json manifest, a README.md, and in some cases implementation code.
Mixed. Some plugins (quant-trading-system, code-safety-monitor, ace-context-engineering) contain implementation code. Most plugins are manifest + README definitions that describe agent prompts and workflows. No centralized test suite.
| Plugin | Category | What It Contains |
|---|---|---|
ace-context-engineering | Productivity | Schemas (7 JSON), scripts (5 Python), skill.md. Implements bullet/delta context evolution with TF-IDF retrieval. |
code-safety-monitor | Safety | Python source in src/, setup.py, requirements.txt. DSPy-based backdoor detection. |
quant-trading-system | Trading | Python modules in qts/ at repo root (main, risk, executor, market_intelligence, tournament, research_loop, llm_client, bullets). |
tournament-runner | Trading | Plugin manifest. Uses qts.tournament module from repo root. |
research-execution-pipeline | Research | Plugin manifest. Uses qts.research_loop from repo root. |
modular-agentic-dev | Methodology | Plugin manifest. Session-based agentic development framework. |
github-automation-suite | GitHub | Plugin manifest. Describes 13 GitHub agent roles. |
swarm-coordination | Distributed | Plugin manifest. Adaptive/hierarchical/mesh swarm patterns. |
consensus-protocols | Distributed | Plugin manifest. Byzantine, Raft, CRDT, gossip protocols. |
core-dev-suite | Development | Plugin manifest. Planner, coder, reviewer, tester, researcher agents. |
market-intelligence | Trading | Plugin manifest. Market analysis agent definitions. |
repo-management-tools | GitHub | Plugin manifest. Multi-repo coordination. |
hive-mind-orchestration | Distributed | Plugin manifest. Distributed decision-making. |
testing-qa-framework | Testing | Plugin manifest. TDD agents and test automation. |
api-docs-generator | Documentation | Plugin manifest. OpenAPI/Swagger generation. |
cicd-automation | DevOps | Plugin manifest. GitHub Actions automation. |
mobile-automation | Mobile | Plugin manifest. React Native + ClassDojo workflows. |
performance-optimization | Optimization | Plugin manifest. Load balancing and benchmarks. |
architecture-design | Architecture | Plugin manifest. System design agents. |
memory-coordination | Coordination | Plugin manifest. Memory and task orchestration. |
safety-compliance | Safety | Plugin manifest. Circuit breakers and kill switches. |
sparc-methodology | Methodology | Plugin manifest. SPARC development framework. |
base-template-generator | Templates | Plugin manifest. Project scaffolding. |
.claude-plugin/
marketplace.json # Plugin registry (23 entries)
plugins/
<plugin-name>/
plugin.json # Manifest (name, description, version)
README.md # Plugin documentation
src/ # (some plugins) Implementation code
qts/ # Quantitative trading system (shared module)
main.py # Entry point
risk.py # Risk management (leverage caps, stop-loss, drawdown)
executor.py # Order execution (paper + live via CCXT)
market_intelligence.py # Market data analysis
tournament.py # Multi-variant paper trading tournaments
research_loop.py # Strategy research pipeline
llm_client.py # LLM provider integration (mock, OpenAI, DeepSeek)
bullets.py # ACE bullet logging
# Add marketplace to Claude Code
/plugin marketplace add jmanhype/claude-code-plugins
# Browse plugins
/plugin
# Install a specific plugin
/plugin install code-safety-monitor
The qts/ directory at the repo root contains the actual trading implementation shared by 3 plugins (quant-trading-system, tournament-runner, research-execution-pipeline).
# Paper trading
PYTHONPATH=. python -m qts.main --symbols ETH SOL XRP --llm-provider mock
# Tournament (multi-variant comparison)
PYTHONPATH=. python -m qts.tournament --variants tp_trail_1.0 tp_trail_1.5 --days 7
Risk defaults: 1.0x gross leverage cap, 25% max position per symbol, -5% daily loss limit, -10% max drawdown halt.
Trading plugins require environment variables:
EXCHANGE_API_KEY=...
EXCHANGE_SECRET=...
PAPER_TRADING=true # default; live trading requires 3-layer approval
GitHub plugins require:
GITHUB_TOKEN=...
/plugin command infrastructure depends on Claude Code's plugin system, which is not publicly documentedqts/ trading system has 1 test file (test_risk.py)ace-context-engineering has a version field in the manifestmarketplace.json) must be manually kept in sync with the plugins/ directory