{"id":37415862,"url":"https://github.com/eps-ai-solutions/claudecli","last_synced_at":"2026-01-16T06:01:19.703Z","repository":{"id":332225862,"uuid":"1133123376","full_name":"EPS-AI-SOLUTIONS/ClaudeCLI","owner":"EPS-AI-SOLUTIONS","description":"HYDRA 10.0 - Advanced AI System with Self-Correction, Few-Shot Learning, Speculative Decoding, Load Balancing \u0026 Semantic RAG","archived":false,"fork":false,"pushed_at":"2026-01-14T16:11:27.000Z","size":978,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"master","last_synced_at":"2026-01-14T16:59:38.363Z","etag":null,"topics":["ai","automation","claude","few-shot-learning","llm","mcp","ollama","powershell","self-correction","speculative-decoding"],"latest_commit_sha":null,"homepage":null,"language":"PowerShell","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/EPS-AI-SOLUTIONS.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-01-12T23:10:15.000Z","updated_at":"2026-01-14T16:11:31.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/EPS-AI-SOLUTIONS/ClaudeCLI","commit_stats":null,"previous_names":["pawelserkowski-lang/claudecli","eps-ai-solutions/claudecli"],"tags_count":2,"template":false,"template_full_name":null,"purl":"pkg:github/EPS-AI-SOLUTIONS/ClaudeCLI","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/EPS-AI-SOLUTIONS%2FClaudeCLI","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/EPS-AI-SOLUTIONS%2FClaudeCLI/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/EPS-AI-SOLUTIONS%2FClaudeCLI/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/EPS-AI-SOLUTIONS%2FClaudeCLI/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/EPS-AI-SOLUTIONS","download_url":"https://codeload.github.com/EPS-AI-SOLUTIONS/ClaudeCLI/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/EPS-AI-SOLUTIONS%2FClaudeCLI/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":28477588,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-01-16T03:13:13.607Z","status":"ssl_error","status_checked_at":"2026-01-16T03:11:47.863Z","response_time":107,"last_error":"SSL_read: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["ai","automation","claude","few-shot-learning","llm","mcp","ollama","powershell","self-correction","speculative-decoding"],"created_at":"2026-01-16T06:00:35.209Z","updated_at":"2026-01-16T06:01:19.656Z","avatar_url":"https://github.com/EPS-AI-SOLUTIONS.png","language":"PowerShell","readme":"# HYDRA 10.0 - ClaudeCLI\n\n[![GitHub stars](https://img.shields.io/github/stars/pawelserkowski-lang/ClaudeCLI?style=flat-square)](https://github.com/pawelserkowski-lang/ClaudeCLI/stargazers)\n[![GitHub forks](https://img.shields.io/github/forks/pawelserkowski-lang/ClaudeCLI?style=flat-square)](https://github.com/pawelserkowski-lang/ClaudeCLI/network/members)\n[![GitHub issues](https://img.shields.io/github/issues/pawelserkowski-lang/ClaudeCLI?style=flat-square)](https://github.com/pawelserkowski-lang/ClaudeCLI/issues)\n[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg?style=flat-square)](https://opensource.org/licenses/MIT)\n[![PowerShell](https://img.shields.io/badge/PowerShell-7.0+-5391FE?style=flat-square\u0026logo=powershell\u0026logoColor=white)](https://github.com/PowerShell/PowerShell)\n[![Windows](https://img.shields.io/badge/Windows-11-0078D6?style=flat-square\u0026logo=windows\u0026logoColor=white)](https://www.microsoft.com/windows)\n[![Ollama](https://img.shields.io/badge/Ollama-Local_AI-000000?style=flat-square)](https://ollama.ai)\n[![Claude](https://img.shields.io/badge/Claude-Anthropic-CC785C?style=flat-square)](https://anthropic.com)\n\n**Maximum Autonomy Mode** | **Parallel Execution** | **MCP Orchestration** | **Advanced AI**\n\n```\n _   ___   ______  ____   ___\n| | | \\ \\ / /  _ \\|  _ \\ / \\ \\\n| |_| |\\ V /| | | | |_) / _ \\ \\\n|  _  | | | | |_| |  _ / ___ \\ \\\n|_| |_| |_| |____/|_| /_/   \\_\\_\\\n\nThree Heads, One Goal. Hydra Executes In Parallel.\n```\n\n## 🎯 Overview\n\nHYDRA 10.0 to zaawansowane środowisko dla **Claude CLI** działające w trybie **Maximum Autonomy Mode** z pełnym dostępem do:\n- Zmiennych środowiskowych (read/write/delete)\n- Rejestru Windows (HKLM, HKCU)\n- Systemu plików (pełen dostęp)\n- Operacji sieciowych (firewall, port scanning)\n- Instalacji oprogramowania (chocolatey, winget, npm, pip)\n\n## 📋 Features\n\n### ✨ Maximum Autonomy Mode\n- 🔓 **Wildcard permissions**: `Bash(*)`, `Write(*)`, `Edit(*)`, `Read(*)`\n- 🔧 **Rejestr Windows**: Pełny R/W dostęp\n- 🔐 **Uprawnienia admin**: RunAs, zarządzanie usługami\n- 📂 **System plików**: Dostęp do System32, Program Files\n- 🌐 **Operacje sieciowe**: Port scanning, firewall rules\n- 📦 **Instalacja software**: choco, winget, npm global, pip\n- ⚡ **Wykonywanie skryptów**: Unrestricted Execution Policy\n\n### ⚡ Parallel Execution\n- Wszystkie niezależne operacje wykonywane równolegle\n- PowerShell Jobs dla MCP health checks\n- Zgodność z zasadą: *\"Każda operacja, która może być wykonana równolegle, MUSI być wykonana równolegle\"*\n\n### 🧠 Advanced AI System (5 Modules)\n\n| Module | Description | Key Feature |\n|--------|-------------|-------------|\n| **Self-Correction** | Auto-validates code with phi3:mini | Regenerates on syntax errors |\n| **Few-Shot Learning** | Learns from successful responses | Context-aware examples |\n| **Speculative Decoding** | Parallel multi-model generation | Model racing \u0026 consensus |\n| **Load Balancing** | CPU-aware provider switching | Auto local/cloud selection |\n| **Semantic File Mapping** | Deep RAG with import analysis | Dependency graph context |\n\n```powershell\n# Quick AI commands\nInvoke-AdvancedAI \"Write Python sort\" -Mode code    # Self-correction + few-shot\nGet-AIQuick \"Capital of France?\"                     # Model racing (~2s)\nNew-AICode \"Download file function\"                  # Code with validation\nInvoke-SemanticQuery -FilePath \"app.py\" -Query \"How does auth work?\"\n```\n\n### 📊 AI Health Dashboard\n- Podgląd stanu providerów, tokenów i kosztów\n- Tryb JSON do integracji z monitoringiem\n\n```powershell\n.\\ai-handler\\Invoke-AIHealth.ps1\n.\\ai-handler\\Invoke-AIHealth.ps1 -Json\n```\n\n### 🔐 Szyfrowanie danych\n- Stan AI i kolejki są szyfrowane AES-256\n- Klucz: `CLAUDECLI_ENCRYPTION_KEY` w zmiennych środowiskowych\n\n### 🛠️ MCP Tools Integration\n| Tool | Port | Transport | Funkcja |\n|------|------|-----------|---------|\n| **Serena** | 9000 | SSE | Symbolic code analysis |\n| **Desktop Commander** | 8100 | Stdio | System operations |\n| **Playwright** | 5200 | Stdio | Browser automation |\n\n## 📦 Installation\n\n### Prerequisites\n- Windows 11\n- PowerShell 7+ (pełne funkcje: streaming, równoległość); 5.1 działa dla podstawowych komend\n- Claude CLI\n- API Key: `ANTHROPIC_API_KEY`\n- Encryption Key: `CLAUDECLI_ENCRYPTION_KEY`\n\n### Quick Start\n\n```powershell\n# 1. Clone or download projekt do Desktop\\ClaudeCLI\n\n# 2. Ustaw API Key\n[Environment]::SetEnvironmentVariable('ANTHROPIC_API_KEY', 'sk-ant-api03-...', 'User')\n[Environment]::SetEnvironmentVariable('OPENAI_API_KEY', 'sk-...', 'User')\n[Environment]::SetEnvironmentVariable('GOOGLE_API_KEY', '...', 'User')\n[Environment]::SetEnvironmentVariable('MISTRAL_API_KEY', '...', 'User')\n[Environment]::SetEnvironmentVariable('GROQ_API_KEY', '...', 'User')\n[Environment]::SetEnvironmentVariable('CLAUDECLI_ENCRYPTION_KEY', '...', 'User')\n\n# 3. Uruchom launcher\n.\\ClaudeCLI.vbs\n# LUB\npowershell -ExecutionPolicy Bypass -File _launcher.ps1\n\n# 4. (Opcjonalnie) Utwórz shortcut na pulpicie\n.\\create-shortcuts.ps1\n```\n\n## 🗂️ Project Structure\n\n```\nC:\\Users\\BIURODOM\\Desktop\\ClaudeCLI\\\n├── .claude/\n│   ├── commands/            # Custom slash commands (ai, ai-batch, ai-config...)\n│   ├── hooks/               # Event hooks\n│   ├── skills/              # Custom skills (serena-commander, hydra)\n│   └── statusline.js        # Status bar config\n├── .serena/\n│   └── project.yml          # Serena project config\n├── ai-handler/              # 🤖 AI Model Handler\n│   ├── AIModelHandler.psm1  # Main module\n│   ├── ai-config.json       # Provider/model configuration\n│   ├── Invoke-AIHealth.ps1  # Health dashboard\n│   ├── modules/             # 🧠 Advanced AI Modules\n│   │   ├── SelfCorrection.psm1\n│   │   ├── FewShotLearning.psm1\n│   │   ├── SpeculativeDecoding.psm1\n│   │   ├── LoadBalancer.psm1\n│   │   ├── SemanticFileMapping.psm1\n│   │   └── AdvancedAI.psm1\n│   └── *.ps1                # CLI wrappers \u0026 tests\n├── parallel/                # ⚡ Parallel execution system\n│   ├── modules/ParallelUtils.psm1\n│   └── scripts/             # Git, Download, Compress, TaskDAG...\n├── CLAUDE.md                # System instructions (900+ lines)\n├── README.md                # This file\n├── _launcher.ps1            # Main launcher\n├── mcp-servers.json         # MCP server configuration\n└── mcp-health-check.ps1     # MCP diagnostics\n```\n\n## 🔐 Security Policy\n\n### Environment Variables Access\nClaudeCLI ma **pełny dostęp** do zmiennych środowiskowych:\n\n```powershell\n# ✅ DOZWOLONE: Odczyt\n$apiKey = $env:ANTHROPIC_API_KEY\n\n# ✅ DOZWOLONE: Modyfikacja (User/Machine/Process scopes)\n[Environment]::SetEnvironmentVariable('NEW_VAR', 'value', 'User')\n\n# ✅ DOZWOLONE: Usuwanie\n[Environment]::SetEnvironmentVariable('OLD_VAR', $null, 'User')\n```\n\n### Permissions Model\n\n```json\n{\n  \"permissions\": {\n    \"allow\": [\n      \"Bash(*)\",                    // WSZYSTKIE komendy Bash\n      \"mcp__serena__*\",\n      \"mcp__desktop-commander__*\",\n      \"mcp__playwright__*\",\n      \"Write(*)\", \"Edit(*)\", \"Read(*)\",\n      \"Glob(*)\", \"Grep(*)\",\n      \"Skill(*)\", \"SlashCommand(*)\"\n    ],\n    \"deny\": []                      // Pusta lista\n  }\n}\n```\n\n### 🚨 Absolutne zakazy (nawet w Maximum Autonomy):\n- `rm -rf /` lub `Remove-Item C:\\ -Recurse -Force`\n- `format C:`\n- `diskpart` bez potwierdzenia\n- Masowe usuwanie kluczy rejestru\n- Wyłączanie Windows Defender bez zgody\n\n**Filozofia**: *Pełna moc, ale z odpowiedzialnością. Przed destrukcyjnymi operacjami - pytaj użytkownika!*\n\n## 🚀 Usage\n\n### Podstawowe uruchomienie\n\n```powershell\n# Via VBS launcher (zalecane)\n.\\ClaudeCLI.vbs\n\n# Via PowerShell\npowershell -ExecutionPolicy Bypass -File _launcher.ps1\n```\n\n### MCP Health Check\n\n```powershell\n# Ręczne sprawdzenie MCP servers (parallel execution)\n.\\mcp-health-check.ps1 -TimeoutSeconds 5\n\n# Skrypt zawsze inicjalizuje AI Handler przy starcie (banner + dostępne modele)\n\n# Przykłady rozszerzonych opcji\n.\\mcp-health-check.ps1 -Server Serena -HostName 127.0.0.1 -RetryCount 3\n.\\mcp-health-check.ps1 -Json -ExportJsonPath .\\\\logs\\\\health.json\n.\\mcp-health-check.ps1 -NoColor -ExportCsvPath .\\\\logs\\\\health.csv\n.\\mcp-health-check.ps1 -AutoRestart\n```\n\n### Konfiguracja MCP servers\n- Edytuj `mcp-servers.json`, aby dodać/zmienić serwery MCP bez dotykania skryptu.\n\n### Tworzenie skrótu na pulpicie\n\n```powershell\n.\\create-shortcuts.ps1\n```\n\n## 🔧 Configuration\n\n### API Keys Setup\n\n```powershell\n# PowerShell (User scope - rekomendowane)\n[Environment]::SetEnvironmentVariable('ANTHROPIC_API_KEY', 'sk-ant-api03-...', 'User')\n\n# PowerShell (Process scope - tymczasowe)\n$env:ANTHROPIC_API_KEY = \"sk-ant-api03-...\"\n\n# CMD\nsetx ANTHROPIC_API_KEY \"sk-ant-api03-...\"\n```\n\n### Dodatkowa konfiguracja\n\n```powershell\n# (Opcjonalnie) Nadpisanie katalogu projektu\n$env:CLAUDECLI_ROOT = \"C:\\\\Users\\\\%USERNAME%\\\\Desktop\\\\ClaudeCLI\"\n```\n\n### Weryfikacja\n\n```powershell\n# Lista zmiennych Claude/Anthropic\nGet-ChildItem env: | Where-Object { $_.Name -like \"*CLAUDE*\" -or $_.Name -like \"*ANTHROPIC*\" }\n\n# Sprawdź konkretną zmienną (maskowana)\n$key = $env:ANTHROPIC_API_KEY\nif ($key) { Write-Host \"✓ API Key: $($key.Substring(0,15))...\" }\n```\n\n## 📚 Architecture\n\n### Council of Six (Multi-Agent Debate)\n\n| Agent | Rola | Fokus |\n|-------|------|-------|\n| **Architekt** | Fakty | Rust 2024, React 19, czysta struktura |\n| **Security** | Ryzyko | ENV vars allowed, zero commits wrażliwych danych, maskowanie kluczy API |\n| **Speedster** | Performance | Lighthouse \u003e 90, bundle \u003c 200KB |\n| **Pragmatyk** | Korzyści | Hybrydowość Web + Desktop |\n| **Researcher** | Weryfikacja | Sprawdzaj w docs/Google przed implementacją |\n| **Jester** | Emocje | Krytyka boilerplate'u i over-engineeringu |\n\n### Parallel Execution Principle\n\n```rust\n// DOBRZE: tokio::join! dla niezależnych operacji\nlet (a, b, c) = tokio::join!(task_a(), task_b(), task_c());\n\n// ŹLE: sekwencyjne await\nlet a = task_a().await;\nlet b = task_b().await;  // marnowanie czasu\n```\n\n```typescript\n// DOBRZE: Promise.all\nconst [users, products] = await Promise.all([fetchUsers(), fetchProducts()]);\n\n// ŹLE: await waterfall\nconst users = await fetchUsers();\nconst products = await fetchProducts();\n```\n\n## 🛡️ Best Practices\n\n| Zalecenie | Priorytet |\n|-----------|-----------|\n| Używaj ENV vars zamiast hardcoded keys | 🔴 Krytyczny |\n| NIE commituj kluczy API do Git | 🔴 Krytyczny |\n| Maskuj klucze API w outputach (15 znaków) | 🟡 Średni |\n| Preferuj absolute paths | 🟢 Niski |\n| Zawsze używaj error handling (try/catch) | 🟡 Średni |\n| Parallel MCP calls gdy możliwe | 🟡 Średni |\n| Backupy przed systemowymi zmianami | 🟡 Średni |\n\n## 📖 Documentation\n\nPełna dokumentacja systemowa: **[CLAUDE.md](CLAUDE.md)** (386 linii)\n\nZawiera:\n- Parallel Execution (Zasada Nadrzędna)\n- Council of Six (Multi-Agent Debate)\n- Tech Stack\n- Project Structure\n- Security Policy (Maximum Autonomy Mode)\n- Protocols (PowerShell, MCP)\n- Best Practices\n\n## 🤝 Contributing\n\nTen projekt działa w trybie **Maximum Autonomy**. Przed wprowadzeniem zmian:\n\n1. Przeczytaj **CLAUDE.md** (instrukcje systemowe)\n2. Przestrzegaj zasad **Parallel Execution**\n3. Używaj **try/catch** z `-ErrorAction Stop`\n4. Zawsze **absolute paths**\n5. **Loguj z kolorami** (Write-Host -ForegroundColor)\n\n## 📜 License\n\nMIT License - see [LICENSE](LICENSE) for details\n\n## 🔗 Links\n\n- [Claude CLI Documentation](https://docs.anthropic.com/en/docs/claude-code)\n- [Anthropic API Console](https://console.anthropic.com/)\n- [MCP Servers](https://modelcontextprotocol.io/)\n\n---\n\n\u003e *\"Trzy głowy, jeden cel. Hydra wykonuje równolegle.\"*\n\n**HYDRA 10.0** | Maximum Autonomy Mode | Windows 11\n","funding_links":[],"categories":[],"sub_categories":[],"project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Feps-ai-solutions%2Fclaudecli","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Feps-ai-solutions%2Fclaudecli","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Feps-ai-solutions%2Fclaudecli/lists"}