Skip to main content

Transform your AI coding assistant into a productivity powerhouse with custom tools and workflows

Project description

TASAK: The Agent's Swiss Army Knife

Transform your AI coding assistant into a productivity powerhouse with custom tools and workflows tailored to YOUR codebase.

🚀 Why TASAK?

For AI Agent Power Users (Claude Code, Cursor, Copilot)

Problem: Your AI assistant wastes tokens rediscovering your project structure, can't run your custom toolchain, and you're copy-pasting commands back and forth.

Solution: TASAK gives your AI agent a curated toolkit that understands YOUR workflow:

  • 📦 Package complex workflows into simple commands ("deploy staging" instead of 10 manual steps)
  • 🧠 Reduce context usage by 80% through hierarchical command discovery
  • 🔧 Self-improving: Let your agent write Python plugins to extend its own capabilities!
  • 🎯 Project-aware: Different tools for different projects, automatically

For Development Teams

Problem: Every developer has their own way of running tests, deployments, and dev environments. Onboarding is painful.

Solution: TASAK standardizes your team's workflow into a unified command palette:

  • 🏢 Company-wide tooling in global config, project-specific in local
  • 📚 Self-documenting: Your AI agent can explain and execute any workflow
  • 🔒 Secure by default: Only expose what you explicitly allow
  • 🚄 Zero friction: Works with any language, any framework, any toolchain

💡 Real-World Magic

# Your AI agent can now do THIS with a single command:
tasak deploy_review_app
# Instead of:
# 1. Check git branch
# 2. Build Docker image
# 3. Push to registry
# 4. Update k8s manifests
# 5. Apply to cluster
# 6. Wait for rollout
# 7. Run smoke tests
# 8. Post PR comment with URL

🎯 Perfect For

✨ Claude Code / Cursor / Copilot / Gemini CLI / Codex CLI / Users

  • Build a custom toolkit that makes your AI assistant 10x more effective
  • Stop wasting time on repetitive commands - let your agent handle them
  • Create project-specific "skills" your AI can use intelligently

👥 Development Teams

  • Standardize workflows across your entire team
  • Make complex operations accessible to junior developers
  • Document-by-doing: your commands ARE the documentation

🔧 DevOps & Platform Engineers

  • Expose safe, curated access to production tools
  • Build guardrails around dangerous operations
  • Create approval workflows for sensitive commands

🎨 Open Source Maintainers

  • Give contributors a standard way to run your project
  • Reduce "works on my machine" issues
  • Make your project AI-assistant friendly

🌟 Killer Features

🧩 Python Plugins (NEW!)

Your AI agent can write its own tools! Just ask:

"Create a plugin that formats all Python files and runs tests"

The agent writes the Python function, TASAK automatically loads it. Mind = blown. 🤯

🎭 Three Modes of Power

cmd apps - Quick & dirty commands

format_code:
  type: cmd
  meta:
    command: "ruff format . && ruff check --fix"

mcp apps - Stateful AI-native services

database:
  type: mcp
  meta:
    command: "uvx mcp-server-sqlite --db ./app.db"

curated apps - Orchestrated workflows

full_deploy:
  type: curated
  commands:
    - test
    - build
    - deploy
    - notify_slack

🔄 Hierarchical Config

Global tools + project tools = perfect setup

~/.tasak/tasak.yaml       # Your personal toolkit
./project/tasak.yaml      # Project-specific tools
= Your AI has exactly what it needs

⚡ Quick Start

1. Install (30 seconds)

pipx install git+https://github.com/jacekjursza/TASAK.git

2. Create Your First Power Tool (1 minute)

cat > ~/.tasak/tasak.yaml << 'EOF'
header: "My AI Assistant Toolkit"

apps_config:
  enabled_apps:
    - dev
    - test
    - deploy

# One command to rule them all
dev:
  name: "Start Development"
  type: "cmd"
  meta:
    command: "docker-compose up -d && npm run dev"

test:
  name: "Run Tests"
  type: "cmd"
  meta:
    command: "npm test && npm run e2e"

deploy:
  name: "Deploy to Staging"
  type: "cmd"
  meta:
    command: "./scripts/deploy.sh staging"
EOF

3. Watch Your AI Agent Level Up

# Your AI can now:
tasak dev      # Start entire dev environment
tasak test     # Run full test suite
tasak deploy   # Deploy to staging
# No more copy-pasting commands!

🎓 Real Use Cases

Use Case 1: Supercharge Your Claude Code

# .tasak/tasak.yaml in your project
header: "NextJS + Supabase Project"

apps_config:
  enabled_apps:
    - setup_branch
    - check_types
    - preview

setup_branch:
  name: "Setup new feature branch"
  type: "cmd"
  meta:
    command: |
      git checkout -b $1 &&
      npm install &&
      npm run db:migrate &&
      npm run dev

check_types:
  name: "Full type check"
  type: "cmd"
  meta:
    command: "tsc --noEmit && eslint . --fix"

preview:
  name: "Deploy preview"
  type: "cmd"
  meta:
    command: "vercel --prod=false"

Now your Claude Code can:

  • Create and setup feature branches
  • Run comprehensive type checks
  • Deploy preview environments ...all without you typing a single command!

Use Case 2: Team Workflow Standardization

# Company-wide ~/.tasak/tasak.yaml
header: "ACME Corp Standard Tools"

apps_config:
  enabled_apps:
    - vpn
    - staging_logs
    - prod_deploy

vpn:
  name: "Connect to VPN"
  type: "cmd"
  meta:
    command: "openvpn --config ~/.acme/vpn.conf"

staging_logs:
  name: "Stream staging logs"
  type: "cmd"
  meta:
    command: "kubectl logs -f -n staging --selector=app"

prod_deploy:
  name: "Production deployment"
  type: "curated"
  commands:
    - name: "deploy"
      description: "Full production deployment with approvals"
      backend:
        type: composite
        steps:
          - type: cmd
            command: ["./scripts/request-approval.sh"]
          - type: cmd
            command: ["./scripts/deploy-prod.sh"]

Use Case 3: Python Plugins - Let AI Extend Itself!

# Your AI agent can write this!
# ~/.tasak/plugins/my_tools.py

def smart_refactor(file_pattern: str, old_name: str, new_name: str):
    """Refactor variable/function names across multiple files"""
    import subprocess
    result = subprocess.run(
        ["rg", "-l", old_name, file_pattern],
        capture_output=True,
        text=True
    )
    files = result.stdout.strip().split("\n")

    for file in files:
        subprocess.run([
            "sed", "-i", f"s/{old_name}/{new_name}/g", file
        ])

    return f"Refactored {len(files)} files"

# Now available as: tasak smart_refactor "*.py" "oldFunc" "newFunc"

📚 Documentation

Quick Links:

🤝 Community & Support

🛠️ For Contributors

Built with Python 3.11+, following TDD principles. We welcome contributions!

Development Setup

git clone https://github.com/jacekjursza/TASAK.git
cd TASAK
python -m venv .venv
source .venv/bin/activate
pip install -e .
pytest -q  # Run tests

See CONTRIBUTING.md for guidelines.

📄 License

MIT License - see LICENSE for details.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

tasak-0.1.0.tar.gz (71.1 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

tasak-0.1.0-py3-none-any.whl (52.9 kB view details)

Uploaded Python 3

File details

Details for the file tasak-0.1.0.tar.gz.

File metadata

  • Download URL: tasak-0.1.0.tar.gz
  • Upload date:
  • Size: 71.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for tasak-0.1.0.tar.gz
Algorithm Hash digest
SHA256 57932b8264910c6abd8ef7856088b6218defdc6adc9a40fdbbb3d7a51a6dc025
MD5 1497d51396ee6096619aa78c51ad4998
BLAKE2b-256 e700a4f83dfd030e159705cc158c6ee6edbb22652ba28f38c290fbd1651ada92

See more details on using hashes here.

Provenance

The following attestation bundles were made for tasak-0.1.0.tar.gz:

Publisher: publish.yml on jacekjursza/TASAK

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file tasak-0.1.0-py3-none-any.whl.

File metadata

  • Download URL: tasak-0.1.0-py3-none-any.whl
  • Upload date:
  • Size: 52.9 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for tasak-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 8751f6e8e8af858cfc47368128294a3b1800b2b6f39d48087b096bfc7513dc8f
MD5 369eb101b57a4051ec7aed4c907975dd
BLAKE2b-256 a2803f9bc81ce2b117a3c2e16e080e1716fc144decd4d59cac8df90569e02bee

See more details on using hashes here.

Provenance

The following attestation bundles were made for tasak-0.1.0-py3-none-any.whl:

Publisher: publish.yml on jacekjursza/TASAK

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page