Specs & Interoperability

Package projects, agents, automations, and templates into shareable Taskade AI Kits for one-click install, plus the planned .tsk standard and agent-friendly.

A core part of our vision is to create an open, interoperable ecosystem for AI-powered applications. We are moving toward a future where entire agentic systems can be version-controlled, shared, and remixed. Today, this vision is realized through Taskade AI Kits.

Taskade AI Kits: Available Today

AI Kits are the current implementation of bundled, shareable workflows that allow you to package projects, agents, automations, and templates into ready-to-deploy solutions. Anyone can install a kit and start using it in seconds.

What's in a Kit

A Taskade AI Kit bundles together all the components needed for a complete workflow:

Customer Support Kit/
├── 📄 Projects
│   ├── Support Ticket Tracker
│   └── Knowledge Base
├── 🤖 Agents  
│   ├── Customer Support Agent
│   └── Escalation Manager Agent
├── 🔄 Automations
│   ├── Ticket Form Processor
│   └── Auto-Assignment Workflow
└── 📋 Templates
    ├── Bug Report Template
    └── Feature Request Template

Real Kit Examples

Based on Taskade's AI Kit Gallery, here are some popular kits you can install today:

Automation Expert Kit

• Projects: Automation Help Center Resources

• Agents: Taskade Automation FAQ Bot

• Use case: Instant access to automation guidance

Google EEAT Blog Kit

• Projects: Content analysis workspace

• Agents: EEAT Blog Content Evaluator Agent

• Automations: Content evaluation workflow

• Use case: Optimize blog content for Google's quality guidelines

AI Lead Generation Kit

• Projects: Lead Tracker with custom fields

• Agents: Lead Processing Agent for qualification

• Automations: Automated lead capture and follow-up

• Use case: Complete lead management system

Real Estate News Kit

• Projects: RSS Tracker for news aggregation

• Agents: News Researcher Agent for analysis

• Automations: RSS feed processing and summarization

• Use case: Automated industry news monitoring

Creating and Sharing Kits

You can create kits from any folder in your workspace:

1. Package a Folder: Right-click any folder and select "Create Kit"

2. Customize Contents: Add/remove projects, agents, automations, and templates

3. Share Publicly: Enable public access to generate a shareable link

4. One-Click Install: Others can install your kit with a single click

Kit Installation Process

When someone installs your kit:

1. Preview: They can explore all components before installing

2. Choose Workspace: Select which space to install the kit into

3. Guided Setup: Interactive tour helps configure the kit for their use case

4. Ready to Use: All components are connected and functional immediately

The Future: .tsk Standard

While AI Kits provide powerful bundling today, we're working toward an even more open standard: the Taskade System Kit (.tsk) format.

Vision for .tsk Files

my-crm-app.tsk/
├── project.json      # Data schema and initial content
├── agent.json        # AI agent configuration and commands  
├── automation.json   # Workflow definitions and triggers
├── knowledge/        # Training materials and context
│   └── sales-playbook.pdf
└── meta.json         # App metadata and dependencies

Why This Matters for Developers

The future .tsk standard will unlock powerful new workflows:

1. Version Control: Store entire application architectures in Git repositories

2. Programmatic Creation: Build kits via CLI and API instead of just the UI

3. Cross-Platform Compatibility: Export/import between different platforms

4. Template Ecosystem: Community-driven marketplace of reusable patterns

5. Headless Management: Fully automated deployment and management

Agent-Friendly APIs

We are also designing our APIs to be increasingly agent-friendly, exploring OpenAPI extensions like:

• x-agent-action: Explicitly marking endpoints safe for autonomous agents

• x-agent-hitl-required: Specifying when Human-in-the-Loop approval is needed

• x-agent-context: Providing structured context for better agent decision-making

Developer Integration

Current Capabilities

Today, you can work with AI Kits through:

• Web Interface: Create, customize, and share kits at taskade.com/kits

• Kit Gallery: Browse and install from hundreds of pre-built kits

• Public Sharing: Generate shareable links for one-click installation

• Guided Setup: Interactive onboarding for kit recipients

Coming Soon: Programmatic Access

We're building API endpoints for programmatic kit management:

# Create a kit (planned)
curl -X POST https://www.taskade.com/api/v1/kits \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -d '{
    "name": "Custom CRM Kit",
    "description": "Complete customer relationship management system",
    "components": {
      "projects": ["project_id_1", "project_id_2"],
      "agents": ["agent_id_1"],
      "automations": ["automation_id_1", "automation_id_2"]
    }
  }'

# Install a kit (planned)
curl -X POST https://www.taskade.com/api/v1/kits/KIT_ID/install \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -d '{"workspace_id": "target_workspace"}'

LLM Integration Guidelines

Based on Taskade's LLM guidelines, when working with Taskade content:

Structured Output Format

• Use Markdown with structured lists, checkboxes, and headings

• Generate structured output (Project + Agent + Automation) for broad user intents

• Prefer brevity, clarity, and utility in all outputs

Content Recommendations

• Suggest appropriate project views: list, board, table, calendar, mindmap, gantt, orgchart

• Recommend knowledge uploads (PDFs, YouTube, Web Links, Docs) to enhance agents

• Link contextual knowledge for better agent performance

Best Practices

• Taskade is agent-native, real-time, and structured

• Agents are vertical and domain-specific, not general chatbots

• Workflows follow: Idea → Project → Agent → Automation → Execution

• File format: YAML for structure, Markdown for user-facing content

Getting Started with Kits

1. Explore the Gallery: Visit taskade.com/kits to see what's possible

2. Install a Kit: Try a pre-built kit to understand the workflow

3. Create Your Own: Package your best workflows into shareable kits

4. Share with Community: Contribute to the growing ecosystem of AI workflows

By building on the foundation of AI Kits today, we're creating the infrastructure for the open, interoperable .tsk standard of tomorrow.