Agentic Workflow OS

Structured Architectural Scaffolding for AI Development

Agentic Workflow OS is a development platform that orchestrates Multi-Agent Systems to plan, architect, and implement complex software projects. Unlike "Chat with Code" tools that rely on messy, unstructured conversation history, this system enforces a Context-First philosophy. It treats Agent Context as a file-system state machine, ensuring that your AI Engineer knows exactly what your AI Architect decided.

---

⚡ Why Use This?

Agentic Workflow OS - Because AI deserves structure, and developers deserve results.

Most AI coding tools suffer from Context Drift. After 20 messages, the AI forgets the architectural constraints you set in message #1.

AI for projects and work is like a rocket: it gives a quick boost of velocity if you have direction, but soon everyone realizes it's only a first-stage booster that stops midway once context is off. The velocity drops, and to carry forward and build production-grade applications needs a 2nd and 3rd stage booster with precision guidance to reach and stay in orbit. This project is that 2nd and 3rd stage.

Agentic Workflow OS solves this by:

1. Strict Role Separation: Agents have clearly defined roles with explicit boundaries - architects handle specifications, engineers handle implementation, with enforced separation to prevent scope creep.
2. Artifact-Driven Handoffs: Agents cannot proceed until they receive specific, validated artifacts from predecessors, ensuring dependencies are met before progression.
3. Programmatic Agent Interaction: Agents operate in environments with terminal access, using CLI tools to search project state, log decisions, record handoffs, and progress through workflow stages - enabling structured, auditable AI collaboration without conversational drift.

The system operates on a Stateless Core / Stateful Edge model.

🔄 Workflow Comparison

Feature	Agentic Workflow OS	ChatGPT + Manual	GitHub Copilot	VS Code + Cursor
Context Preservation	✅ File-system state machine	❌ Conversation drift	❌ Session-based	❌ Session-based
Role Separation	✅ Strict agent boundaries	❌ Single AI persona	❌ Single AI persona	❌ Single AI persona
Artifact Validation	✅ Automated quality gates	❌ Manual review	❌ No validation	❌ No validation
Audit Trail	✅ Complete history	❌ Lost in chat	❌ No tracking	❌ No tracking
Multi-Agent Coordination	✅ Orchestrated handoffs	❌ Manual coordination	❌ No coordination	❌ No coordination
Production Ready	✅ Enterprise-grade output	⚠️ Requires significant editing	⚠️ Requires significant editing	⚠️ Requires significant editing

---

🚀 Quick Start

Environment Setup

Create and activate a Python virtual environment:

# Create virtual environment
python3 -m venv myproject-env

# Activate the environment
source myproject-env/bin/activate

# On Windows:
source myproject-env\Scripts\activate

Installation

Option A: Pip (Recommended)

pip install agentic-workstation

# First run will launch an interactive setup wizard
agentic --help

First-Time Setup: On first run, Agentic Workflow OS will launch an interactive setup wizard to configure your workspace and preferences. This creates ~/.config/agentic/config.yaml with your settings.

Option B: From Source (Development)

git clone https://github.com/sujith-eag/agentic_workflow.git
cd agentic_workflow
pip install -e .

Uninstallation

Option A: Pip

pip uninstall agentic-workstation

Option B: Docker

# Remove the Docker image
docker rmi agentic-workstation

# Remove any running containers (if any)
docker rm $(docker ps -aq --filter ancestor=agentic-workstation)

---

🛠️ CLI & TUI Usage

For detailed CLI documentation, see CLI_REFERENCE.md. For interactive usage, use the Text User Interface (TUI) with agentic.

Available Commands

The CLI uses a context-aware design - commands available depend on whether you're in a project directory or not.

Global Commands (from repository root)

• agentic init <name> --workflow <type> --description "desc" - Initialize new project with workflow
• agentic list - List all projects or show details for one
• agentic delete <name> - Permanently delete a project
• agentic workflows - List available workflow definitions
• agentic config - View or edit global configuration
• agentic - Launch TUI in global mode

Project Commands (from within project directory)

• agentic status - Show project status and workflow state
• agentic activate <agent_id> - Activate specific agent session
• agentic handoff --to <agent_id> --artifacts "files" - Record agent handoff with artifacts
• agentic decision --title "Title" --rationale "Reason" - Record project decision
• agentic end - End current workflow session
• agentic feedback --target <agent_id> --severity <level> --summary "Summary" - Record feedback for an agent or artifact
• agentic blocker --title "Title" --description "Desc" --blocked-agents "ids" - Record a blocker that prevents progress
• agentic iteration --trigger "Trigger" --impacted-agents "ids" --description "Desc" - Record an iteration in the development process
• agentic assumption --assumption "Assumption" --rationale "Rationale" - Record an assumption that may affect the project
• agentic list-pending - List pending handoffs for the current project
• agentic list-blockers - List active blockers for the current project

---

🖥️ CLI Interaction for AI use

These assets show the command-line help and workflow interaction screens.

Project Help Screen

Watch the project help screen video

Workflow Help Screen

Watch the workflow help screen video

---

🖼️ TUI for Human Flows

Project Creation Flow

Project Usage Flow

Project Usage Video

Watch the project usage video

---

🏗️ Creating & Running a Project

1. Initialize a Project

Navigate to your empty folder and initialize a project.

# From repository root (global context)
agentic init MySaaS --workflow planning --description "My SaaS application project"

Available workflows: planning, implementation, research, workflow-creation

2. Agent Workflow Orchestration

Agents operate in environments with terminal access, using CLI commands to orchestrate structured workflows:

1. Navigate to Project: cd projects/MySaaS

2. Session Activation: User or Agents run agentic activate A-01 to start their session and receive context.

3. Context Processing: Agents access the generated active_session.md file containing their role, instructions, and project state.

4. Artifact Generation: Agents produce required deliverables and save them to specified artifact paths.

5. Handoff Execution: Agents execute handoff commands to progress the workflow:

   agentic handoff --to A-02 --artifacts "artifacts/project_brief.md,artifacts/requirements.md"
   

   The system validates artifact existence and workflow rules before allowing progression. The from agent is automatically inferred from the active session.

6. State Management: Agents can log decisions, record blockers, and query project status using CLI commands throughout the process.

3. Review Status

View the decision tree and current state.

agentic status

---

📦 Workflows

The system supports pluggable Workflow Manifests with four main workflow types.

1. Planning (15 agents: A-00 to A-14)

• Goal: Turn a 1-sentence idea into a full Tech Spec.
• Key Agents:
  • A-00 Orchestrator & Project Controller
  • A-01 Project Guide & Idea Incubation
  • A-02 Planning & Requirements Analyst
  • A-03 Architecture & Technical Design Analyst
  • A-04 Security & Compliance Scrutineer
  • A-05 Infrastructure & DevOps Planning Architect
  • A-06 Data Architecture & Storage Planning Specialist
  • A-07 API & Integration Planning Specialist
  • A-08 UX & Interaction Planning Specialist
  • A-09 Developer Workflow & Engineering Standards Planner

2. Implementation (10 agents: I-00 to I-05 + specialists)

• Goal: Turn specs into production code.
• Key Agents:
  • I-00 Implementation Orchestrator
  • I-01 Scaffold & DevOps Architect
  • I-02 Backend & Data Engineer
  • I-03 Frontend & UX Engineer
  • I-04 Quality & Validation Specialist
  • I-05 Release & Integration Manager
  • I-DOC Documentation Sprint Agent
  • I-DS Data Store Implementation Specialist
  • I-PERF Performance Optimization Agent
  • I-SEC Security Hardening Auditor

3. Research (12 agents)

• Goal: Conduct thorough research, analysis, and reporting.
• Use Cases: Literature review, data analysis, academic research, market research.

4. Workflow-Creation (9 agents)

• Goal: Create and customize new workflow packages.
• Use Cases: Meta-workflow development, process automation.

Custom Workflows

You can create your own workflows by placing a manifest in:

• Project: ./.agentic/workflows/my-custom-flow/workflow.json
• User global: ~/.config/agentic/workflows/my-custom-flow/workflow.json
• System: Bundled in package (not recommended for custom)

Workflow manifests include workflow.json, agents.json, artifacts.json, etc.

---

🛠️ CLI Commands

Global Commands (from repository root)

# List available workflows
agentic workflows

# Initialize new project
agentic init MyProject --workflow planning --description "My project description"

# Launch TUI (Terminal User Interface)
agentic

# Show CLI help
agentic --help

Project Commands (from within project directory)

# Navigate to project
cd projects/MyProject

# Show current project status
agentic status

# Activate an agent session
agentic activate A-01

# Record handoff between agents (from agent inferred from active session)
agentic handoff --to A-02 --artifacts "requirements.md,architecture.md"

# Record project decision
agentic decision --title "Technology Stack Selection" --rationale "React + Node.js chosen for full-stack consistency"

# End workflow session
agentic end

⚙️ Configuration & Setup

Automatic Setup

On first run, Agentic Workflow OS automatically launches an interactive setup wizard that configures:

• Default workspace (where projects are stored)
• Editor command (for opening files)
• UI preferences and logging levels

Configuration Files

Global Config (~/.config/agentic/config.yaml):

default_workspace: "~/AgenticProjects"
editor_command: "code"
tui_enabled: true
check_updates: true
log_level: "INFO"

Project Config (.agentic/config.yaml in each project):

workflow: planning
strict_mode: true
excluded_paths:
  - "node_modules"
  - ".git"
custom_overrides:
  governance:
    require_reviews: true

Manual Configuration

If you prefer to set up configuration manually or modify existing settings:

1. Create global config directory:

   mkdir -p ~/.config/agentic
   

2. Create config file:

   cat > ~/.config/agentic/config.yaml << 'EOF'
   default_workspace: "~/AgenticProjects"
   editor_command: "code"
   tui_enabled: true
   check_updates: true
   log_level: "INFO"
   EOF
   

3. Create workspace directory:

   mkdir -p ~/AgenticProjects
   

Re-running Setup

To re-run the setup wizard or modify configuration:

# Remove config to trigger setup again
rm ~/.config/agentic/config.yaml
agentic

---

🛠️ Development

This project is built with Python 3.10+.

• Core: src/agentic_workflow
• Manifests: src/agentic_workflow/manifests (Canonical Definitions)
• Templates: src/agentic_workflow/templates (Jinja2 templates)
• Build: pyproject.toml (Hatchling)
• Tests: Not included in deployment

License: MIT