agentex-sdk 0.2.6__tar.gz → 0.2.8__tar.gz

agentex_sdk-0.2.8/.release-please-manifest.json

@@ -0,0 +1,3 @@
+{
+  ".": "0.2.8"
+}

{agentex_sdk-0.2.6 → agentex_sdk-0.2.8}/CHANGELOG.md

@@ -1,5 +1,28 @@
 # Changelog
 
+## 0.2.8 (2025-08-09)
+
+Full Changelog: [v0.2.7...v0.2.8](https://github.com/scaleapi/agentex-python/compare/v0.2.7...v0.2.8)
+
+### Chores
+
+* **internal:** update comment in script ([401f1d7](https://github.com/scaleapi/agentex-python/commit/401f1d79034ecb0b556a26debde79681bc21e8ae))
+* update @stainless-api/prism-cli to v5.15.0 ([4d332d0](https://github.com/scaleapi/agentex-python/commit/4d332d0f77a5a11ca6781a5fc7690ae82653cadb))
+
+## 0.2.7 (2025-08-08)
+
+Full Changelog: [v0.2.6...v0.2.7](https://github.com/scaleapi/agentex-python/compare/v0.2.6...v0.2.7)
+
+### Features
+
+* **api:** api update ([e3d08ba](https://github.com/scaleapi/agentex-python/commit/e3d08baad59346db48e04a394a929d6347dafa07))
+* debug features ([40d8db2](https://github.com/scaleapi/agentex-python/commit/40d8db22dcc8f00a6c78e9bc3e1d036ebd1423b6))
+
+
+### Chores
+
+* **internal:** fix ruff target version ([1b880e1](https://github.com/scaleapi/agentex-python/commit/1b880e1dd81d47bb9df12507f13351611ff6367f))
+
 ## 0.2.6 (2025-08-01)
 
 Full Changelog: [v0.2.5...v0.2.6](https://github.com/scaleapi/agentex-python/compare/v0.2.5...v0.2.6)

{agentex_sdk-0.2.6 → agentex_sdk-0.2.8}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: agentex-sdk
-Version: 0.2.6
+Version: 0.2.8
 Summary: The official Python library for the agentex API
 Project-URL: Homepage, https://github.com/scaleapi/agentex-python
 Project-URL: Repository, https://github.com/scaleapi/agentex-python
@@ -126,6 +126,37 @@ asyncio.run(main())
 
 Functionality between the synchronous and asynchronous clients is otherwise identical.
 
+## Debugging
+
+AgentEx provides built-in debugging support for **temporal projects** during local development.
+
+```bash
+# Basic debugging
+uv run agentex agents run --manifest manifest.yaml --debug-worker
+
+# Wait for debugger to attach before starting
+uv run agentex agents run --manifest manifest.yaml --debug-worker --wait-for-debugger
+
+# Custom debug port
+uv run agentex agents run --manifest manifest.yaml --debug-worker --debug-port 5679
+```
+
+For **VS Code**, add this configuration to `.vscode/launch.json`:
+
+```json
+{
+  "name": "Attach to AgentEx Worker",
+  "type": "debugpy",
+  "request": "attach",
+  "connect": { "host": "localhost", "port": 5678 },
+  "pathMappings": [{ "localRoot": "${workspaceFolder}", "remoteRoot": "." }],
+  "justMyCode": false,
+  "console": "integratedTerminal"
+}
+```
+
+The debug server automatically finds an available port starting from 5678 and prints connection details when starting.
+
 ### With aiohttp
 
 By default, the async client uses `httpx` for HTTP requests. However, for improved concurrency performance you may also use `aiohttp` as the HTTP backend.

{agentex_sdk-0.2.6 → agentex_sdk-0.2.8}/README.md

@@ -67,6 +67,37 @@ asyncio.run(main())
 
 Functionality between the synchronous and asynchronous clients is otherwise identical.
 
+## Debugging
+
+AgentEx provides built-in debugging support for **temporal projects** during local development.
+
+```bash
+# Basic debugging
+uv run agentex agents run --manifest manifest.yaml --debug-worker
+
+# Wait for debugger to attach before starting
+uv run agentex agents run --manifest manifest.yaml --debug-worker --wait-for-debugger
+
+# Custom debug port
+uv run agentex agents run --manifest manifest.yaml --debug-worker --debug-port 5679
+```
+
+For **VS Code**, add this configuration to `.vscode/launch.json`:
+
+```json
+{
+  "name": "Attach to AgentEx Worker",
+  "type": "debugpy",
+  "request": "attach",
+  "connect": { "host": "localhost", "port": 5678 },
+  "pathMappings": [{ "localRoot": "${workspaceFolder}", "remoteRoot": "." }],
+  "justMyCode": false,
+  "console": "integratedTerminal"
+}
+```
+
+The debug server automatically finds an available port starting from 5678 and prints connection details when starting.
+
 ### With aiohttp
 
 By default, the async client uses `httpx` for HTTP requests. However, for improved concurrency performance you may also use `aiohttp` as the HTTP backend.

agentex_sdk-0.2.8/examples/launch-tutorials.sh

@@ -0,0 +1,336 @@
+#!/bin/bash
+
+# AgentEx Tutorial Launcher
+# This script helps you easily launch and test all tutorials in the repository
+# 
+# Usage:
+#   ./launch-tutorials.sh           # Show interactive menu
+#   ./launch-tutorials.sh 1         # Launch tutorial #1 directly  
+#   ./launch-tutorials.sh a         # Launch all tutorials with confirmations
+#   ./launch-tutorials.sh c         # Clean up orphaned tutorial processes
+#
+# Note: Excludes 90_multi_agent_non_temporal (use its own start-agents.sh)
+
+# Simple cleanup function for orphaned processes
+cleanup() {
+    # Kill any remaining agentex or uvicorn processes from tutorials
+    local agentex_pids=$(pgrep -f "agentex agents run.*tutorials" 2>/dev/null || true)
+    if [[ -n "$agentex_pids" ]]; then
+        echo "$agentex_pids" | xargs kill -TERM 2>/dev/null || true
+        sleep 1
+        echo "$agentex_pids" | xargs kill -KILL 2>/dev/null || true
+    fi
+    
+    local uvicorn_pids=$(pgrep -f "uvicorn.*project\." 2>/dev/null || true)
+    if [[ -n "$uvicorn_pids" ]]; then
+        echo "$uvicorn_pids" | xargs kill -TERM 2>/dev/null || true
+        sleep 1
+        echo "$uvicorn_pids" | xargs kill -KILL 2>/dev/null || true
+    fi
+}
+
+# Color codes for output
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+BLUE='\033[0;34m'
+NC='\033[0m' # No Color
+
+# Tutorial definitions
+declare -a TUTORIALS=(
+    "tutorials/00_sync/000_hello_acp|Basic Hello ACP (Sync)"
+    "tutorials/00_sync/010_multiturn|Multi-turn Chat (Sync)"
+    "tutorials/00_sync/020_streaming|Streaming Response (Sync)"
+    "tutorials/10_agentic/00_base/000_hello_acp|Basic Hello ACP (Agentic)"
+    "tutorials/10_agentic/00_base/010_multiturn|Multi-turn Chat (Agentic)"
+    "tutorials/10_agentic/00_base/020_streaming|Streaming Response (Agentic)"
+    "tutorials/10_agentic/00_base/030_tracing|Tracing Example (Agentic)"
+    "tutorials/10_agentic/00_base/040_other_sdks|Other SDKs Integration (Agentic)"
+    "tutorials/10_agentic/00_base/080_batch_events|Batch Events (Agentic)"
+    "tutorials/10_agentic/10_temporal/000_hello_acp|Basic Hello ACP (Temporal)"
+    "tutorials/10_agentic/10_temporal/010_agent_chat|Agent Chat (Temporal)"
+    "tutorials/10_agentic/10_temporal/020_state_machine|State Machine (Temporal)"
+)
+
+# Function to print colored output
+print_colored() {
+    local color=$1
+    local message=$2
+    # Check if terminal supports colors
+    if [[ -t 1 ]] && command -v tput >/dev/null 2>&1; then
+        printf "${color}%s${NC}\n" "$message"
+    else
+        printf "%s\n" "$message"
+    fi
+}
+
+# Function to display the menu
+show_menu() {
+    print_colored $BLUE "╔════════════════════════════════════════════════════════════════╗"
+    print_colored $BLUE "║                    AgentEx Tutorial Launcher                  ║"
+    print_colored $BLUE "╚════════════════════════════════════════════════════════════════╝"
+    echo ""
+    print_colored $YELLOW "Available tutorials:"
+    echo ""
+    
+    local index=1
+    for tutorial in "${TUTORIALS[@]}"; do
+        IFS='|' read -r path description <<< "$tutorial"
+        if [[ -t 1 ]] && command -v tput >/dev/null 2>&1; then
+            printf "${GREEN}%2d.${NC} %s\n" $index "$description"
+        else
+            printf "%2d. %s\n" $index "$description"
+        fi
+        index=$((index + 1))
+    done
+    
+    echo ""
+    print_colored $BLUE "Other options:"
+    print_colored $GREEN "  a. Run all tutorials sequentially (with confirmations)"
+    print_colored $GREEN "  c. Clean up any orphaned tutorial processes"
+    print_colored $GREEN "  q. Quit"
+    echo ""
+    print_colored $YELLOW "📌 Note: The multi-agent system tutorial (tutorials/10_agentic/90_multi_agent_non_temporal) is excluded"
+    print_colored $YELLOW "   as it has a special launch process. Use its own start-agents.sh script."
+    echo ""
+}
+
+# Function to run a specific tutorial
+run_tutorial() {
+    local tutorial_index=$1
+    local tutorial_info="${TUTORIALS[$((tutorial_index - 1))]}"
+    IFS='|' read -r path description <<< "$tutorial_info"
+    
+    local manifest_path="${path}/manifest.yaml"
+    
+    print_colored $BLUE "╔════════════════════════════════════════════════════════════════╗"
+    printf "║ Running: %-54s ║\n" "$description"
+    print_colored $BLUE "╚════════════════════════════════════════════════════════════════╝"
+    
+    if [[ ! -f "$manifest_path" ]]; then
+        print_colored $RED "❌ Error: Manifest file not found at $manifest_path"
+        return 1
+    fi
+    
+    print_colored $YELLOW "📂 Tutorial path: $path"
+    print_colored $YELLOW "📄 Manifest: $manifest_path"
+    echo ""
+    print_colored $GREEN "🚀 Executing: cd .. && uv run agentex agents run --manifest examples/$manifest_path"
+    print_colored $YELLOW "💡 Press Ctrl+C to stop the tutorial"
+    echo ""
+    
+    # Run the tutorial directly (need to go to parent dir where uv project is)
+    (cd .. && uv run agentex agents run --manifest "examples/$manifest_path")
+    
+    local exit_code=$?
+    if [[ $exit_code -eq 0 ]]; then
+        print_colored $GREEN "✅ Tutorial completed successfully!"
+    elif [[ $exit_code -eq 130 ]]; then
+        print_colored $YELLOW "🛑 Tutorial was interrupted by user"
+    else
+        print_colored $RED "❌ Tutorial failed with exit code: $exit_code"
+    fi
+    
+    return $exit_code
+}
+
+# Function to run all tutorials
+run_all_tutorials() {
+    print_colored $BLUE "🎯 Running all tutorials sequentially..."
+    echo ""
+    
+    local success_count=0
+    local total_count=${#TUTORIALS[@]}
+    
+    for i in $(seq 1 $total_count); do
+        local tutorial_info="${TUTORIALS[$((i - 1))]}"
+        IFS='|' read -r path description <<< "$tutorial_info"
+        
+        print_colored $YELLOW "┌─ Tutorial $i/$total_count: $description"
+        echo ""
+        
+        # Ask for confirmation
+        while true; do
+            print_colored $BLUE "Run this tutorial? (y/n/q to quit): "
+            read -r response
+            case $response in
+                [Yy]* ) 
+                    if run_tutorial $i; then
+                        success_count=$((success_count + 1))
+                    fi
+                    break
+                    ;;
+                [Nn]* ) 
+                    print_colored $YELLOW "⏭️  Skipping tutorial $i"
+                    break
+                    ;;
+                [Qq]* ) 
+                    print_colored $YELLOW "🛑 Stopping tutorial run"
+                    echo ""
+                    print_colored $BLUE "📊 Summary: $success_count/$((i-1)) tutorials completed successfully"
+                    return 0
+                    ;;
+                * ) 
+                    print_colored $RED "Please answer y, n, or q."
+                    ;;
+            esac
+        done
+        
+        if [[ $i -lt $total_count ]]; then
+            echo ""
+            print_colored $BLUE "────────────────────────────────────────────────────────────────"
+            echo ""
+        fi
+    done
+    
+    echo ""
+    print_colored $BLUE "🎉 All tutorials completed!"
+    print_colored $BLUE "📊 Summary: $success_count/$total_count tutorials completed successfully"
+}
+
+# Function to manually clean up tutorial processes
+manual_cleanup() {
+    print_colored $BLUE "🧹 Manual cleanup of tutorial processes..."
+    echo ""
+    
+    # Check for running tutorial processes
+    local found_processes=false
+    
+    # Check for agentex processes
+    local agentex_pids=$(pgrep -f "agentex agents run.*tutorials" 2>/dev/null || true)
+    if [[ -n "$agentex_pids" ]]; then
+        found_processes=true
+        print_colored $YELLOW "🔍 Found agentex tutorial processes:"
+        ps -p $agentex_pids -o pid,command 2>/dev/null || true
+        echo ""
+    fi
+    
+    # Check for uvicorn processes
+    local uvicorn_pids=$(pgrep -f "uvicorn.*project\." 2>/dev/null || true)
+    if [[ -n "$uvicorn_pids" ]]; then
+        found_processes=true
+        print_colored $YELLOW "🔍 Found uvicorn tutorial processes:"
+        ps -p $uvicorn_pids -o pid,command 2>/dev/null || true
+        echo ""
+    fi
+    
+    # Check for occupied ports
+    print_colored $YELLOW "🔍 Checking common tutorial ports (8000-8003)..."
+    local port_check=$(lsof -i :8000 -i :8001 -i :8002 -i :8003 2>/dev/null || true)
+    if [[ -n "$port_check" ]]; then
+        found_processes=true
+        echo "$port_check"
+        echo ""
+    fi
+    
+    if [[ "$found_processes" == "false" ]]; then
+        print_colored $GREEN "✅ No tutorial processes found - system is clean!"
+        return 0
+    fi
+    
+    # Ask for confirmation before cleaning
+    while true; do
+        print_colored $BLUE "Kill these processes? (y/n): "
+        read -r response
+        case $response in
+            [Yy]* )
+                print_colored $YELLOW "🧹 Cleaning up..."
+                cleanup
+                print_colored $GREEN "✅ Manual cleanup completed!"
+                break
+                ;;
+            [Nn]* )
+                print_colored $YELLOW "⏭️  Cleanup cancelled"
+                break
+                ;;
+            * )
+                print_colored $RED "Please answer y or n."
+                ;;
+        esac
+    done
+}
+
+# Function to validate tutorial number
+validate_tutorial_number() {
+    local num=$1
+    if [[ ! "$num" =~ ^[0-9]+$ ]] || [[ $num -lt 1 ]] || [[ $num -gt ${#TUTORIALS[@]} ]]; then
+        return 1
+    fi
+    return 0
+}
+
+# Main script logic
+main() {
+    # Check if we're in the right directory
+    if [[ ! -f "../pyproject.toml" ]] || [[ ! -d "tutorials" ]]; then
+        print_colored $RED "❌ Error: This script must be run from the examples directory"
+        print_colored $YELLOW "💡 Current directory: $(pwd)"
+        print_colored $YELLOW "💡 Expected files: ../pyproject.toml, tutorials/"
+        exit 1
+    fi
+    
+    # If a tutorial number is provided as argument
+    if [[ $# -eq 1 ]]; then
+        local tutorial_num=$1
+        
+        if [[ "$tutorial_num" == "a" ]] || [[ "$tutorial_num" == "all" ]]; then
+            run_all_tutorials
+            exit 0
+        elif [[ "$tutorial_num" == "c" ]] || [[ "$tutorial_num" == "cleanup" ]]; then
+            manual_cleanup
+            exit 0
+        fi
+        
+        if validate_tutorial_number "$tutorial_num"; then
+            run_tutorial "$tutorial_num"
+            exit $?
+        else
+            print_colored $RED "❌ Error: Invalid tutorial number '$tutorial_num'"
+            print_colored $YELLOW "💡 Valid range: 1-${#TUTORIALS[@]}"
+            exit 1
+        fi
+    fi
+    
+    # Interactive mode
+    while true; do
+        show_menu
+        print_colored $BLUE "Enter your choice (1-${#TUTORIALS[@]}, a, c, or q): "
+        read -r choice
+        
+        case $choice in
+            [Qq]* )
+                print_colored $YELLOW "👋 Goodbye!"
+                exit 0
+                ;;
+            [Aa]* )
+                echo ""
+                run_all_tutorials
+                echo ""
+                ;;
+            [Cc]* )
+                echo ""
+                manual_cleanup
+                echo ""
+                print_colored $BLUE "Press Enter to continue..."
+                read -r
+                ;;
+            * )
+                if validate_tutorial_number "$choice"; then
+                    echo ""
+                    run_tutorial "$choice"
+                    echo ""
+                    print_colored $BLUE "Press Enter to continue..."
+                    read -r
+                else
+                    print_colored $RED "❌ Invalid choice: '$choice'"
+                    print_colored $YELLOW "💡 Please enter a number between 1 and ${#TUTORIALS[@]}, 'a' for all, 'c' for cleanup, or 'q' to quit"
+                fi
+                ;;
+        esac
+        
+        echo ""
+    done
+}
+
+# Run the main function
+main "$@"

agentex_sdk-0.2.8/examples/tutorials/10_agentic/00_base/090_multi_agent_non_temporal/Dockerfile

@@ -0,0 +1,44 @@
+# syntax=docker/dockerfile:1.3
+FROM python:3.12-slim
+
+COPY --from=ghcr.io/astral-sh/uv:0.6.4 /uv /uvx /bin/
+
+# Install system dependencies
+RUN apt-get update && apt-get install -y \
+    htop \
+    vim \
+    curl \
+    tar \
+    python3-dev \
+    postgresql-client \
+    build-essential \
+    libpq-dev \
+    gcc \
+    cmake \
+    netcat-openbsd \
+    && apt-get clean \
+    && rm -rf /var/lib/apt/lists/*
+
+RUN uv pip install --system --upgrade pip setuptools wheel
+
+ENV UV_HTTP_TIMEOUT=1000
+
+# Copy just the pyproject.toml file to optimize caching
+COPY pyproject.toml /app/pyproject.toml
+WORKDIR /app/
+# Install the required Python packages
+RUN uv pip install --system -e .
+
+# Copy the project code
+COPY project /app/project
+
+WORKDIR /app
+
+# Set environment variables
+ENV PYTHONPATH=/app
+
+ARG AGENT_FILE
+ARG PORT
+
+# Run the agent using uvicorn
+CMD uvicorn project.${AGENT_FILE%.*}:acp --host 0.0.0.0 --port ${PORT:-8000}

agentex_sdk-0.2.8/examples/tutorials/10_agentic/00_base/090_multi_agent_non_temporal/README.md

@@ -0,0 +1,197 @@
+# Multi-Agent Content Assembly Line
+
+A multi-agent system that creates content through a collaborative workflow. Four agents work together: a creator generates content, a critic reviews it against rules, and a formatter outputs the final result, all coordinated by an orchestrator.
+
+## 🏗️ Architecture Overview
+
+```
+090_multi_agent_non_temporal/
+├── project/                  # All agent code
+│   ├── creator.py            # Content generation agent
+│   ├── critic.py             # Content review agent  
+│   ├── formatter.py          # Content formatting agent
+│   ├── orchestrator.py       # Workflow coordination agent
+│   ├── models.py             # Pydantic models for type safety
+│   └── state_machines/
+│       └── content_workflow.py  # State machine definitions
+├── creator.yaml              # Creator agent manifest
+├── critic.yaml               # Critic agent manifest
+├── formatter.yaml            # Formatter agent manifest
+├── orchestrator.yaml         # Orchestrator agent manifest
+├── Dockerfile                # Single shared Dockerfile
+├── pyproject.toml            # Dependencies and project configuration
+├── start-agents.sh          # Agent management script
+└── README.md                # This file
+```
+
+## 📁 File Structure
+
+The system uses a shared build configuration with type-safe interfaces:
+- **Single `Dockerfile`** with build arguments for different agents
+- **Single `pyproject.toml`** for all dependencies  
+- **Agent code** in `project/` directory with clear separation of concerns
+- **Individual manifest files** at root level for each agent deployment
+- **Shared state machine definitions** for workflow coordination
+- **Pydantic models** (`models.py`) for type safety and validation across all agents
+
+### Key Files:
+- `project/models.py` - Defines request/response models for type safety
+- `project/orchestrator.py` - Workflow coordination and inter-agent communication
+- `project/creator.py` - Content generation with revision capabilities
+- `project/critic.py` - Content validation against rules
+- `project/formatter.py` - Multi-format content transformation
+- `project/state_machines/content_workflow.py` - State management for the workflow
+
+## 🚀 Quick Start
+
+### Prerequisites
+- Python 3.12+
+- uv package manager
+- OpenAI API key (set `OPENAI_API_KEY` or create `.env` file)
+
+### Running the System
+
+1. **Start all agents**:
+   ```bash
+   cd examples/tutorials/10_agentic/00_base/090_multi_agent_non_temporal
+   ./start-agents.sh start
+   ```
+
+2. **Check agent status**:
+   ```bash
+   ./start-agents.sh status
+   ```
+
+3. **Send a test request**:
+   ```bash
+   ./start-agents.sh test
+   ```
+
+4. **Monitor logs**:
+   ```bash
+   ./start-agents.sh logs
+   ```
+
+5. **Stop all agents**:
+   ```bash
+   ./start-agents.sh stop
+   ```
+
+## 🤖 Agent Responsibilities
+
+### **Creator Agent** (Port 8001)
+- Generates original content based on user requests
+- Revises content based on critic feedback
+- Maintains conversation history and iteration tracking
+
+### **Critic Agent** (Port 8002)  
+- Reviews content against specified rules
+- Provides specific, actionable feedback
+- Approves content when all rules are met
+
+### **Formatter Agent** (Port 8003)
+- Converts approved content to target formats (HTML, Markdown, JSON, etc.)
+- Preserves meaning while applying format-specific conventions
+- Supports multiple output formats
+
+### **Orchestrator Agent** (Port 8000)
+- Coordinates the entire workflow using state machines
+- Manages inter-agent communication
+- Tracks progress and handles errors/retries
+
+## 📋 Example Request
+
+Send a JSON request to the orchestrator:
+
+```json
+{
+  "request": "Write a welcome message for our AI assistant",
+  "rules": ["Under 50 words", "Friendly tone", "Include emoji"],
+  "target_format": "HTML"
+}
+```
+
+The system will:
+1. **Create** content using the Creator agent
+2. **Review** against rules using the Critic agent
+3. **Revise** if needed (up to 10 iterations)
+4. **Format** final approved content using the Formatter agent
+
+## 🔧 Development
+
+### Type Safety with Pydantic
+The tutorial demonstrates proper type safety using Pydantic models:
+
+```python
+# Define request structure
+class CreatorRequest(BaseModel):
+    request: str = Field(..., description="The content creation request")
+    current_draft: Optional[str] = Field(default=None, description="Current draft for revision")
+    feedback: Optional[List[str]] = Field(default=None, description="Feedback from critic")
+
+# Validate incoming requests
+creator_request = CreatorRequest.model_validate(request_data)
+```
+
+Benefits:
+- **Explicit failures** when required fields are missing
+- **Self-documenting** APIs with field descriptions
+- **IDE support** with auto-completion and type checking
+- **Runtime validation** with clear error messages
+
+### Adding New Agents
+1. **Add models** to `project/models.py` for request/response types
+2. **Create agent** in `project/new_agent.py` using the FastACP pattern
+3. **Add manifest** as `new_agent.yaml` at root level with deployment configuration
+4. **Update startup script** in `start-agents.sh` to include the new agent
+
+### Modifying Agents
+- **Agent code** is in `project/` directory
+- **Shared models** are in `project/models.py` for consistency
+- **Dependencies** go in `pyproject.toml`
+- **Docker configuration** is shared across all agents
+
+### Deployment
+Each agent can be deployed independently using its manifest:
+```bash
+uv run agentex agents deploy --cluster your-cluster --manifest creator.yaml
+```
+
+## 🏗️ Technical Implementation
+
+### Shared Dockerfile
+The Dockerfile uses build arguments to run different agents:
+```dockerfile
+CMD uvicorn project.${AGENT_FILE%.*}:acp --host 0.0.0.0 --port ${PORT:-8000}
+```
+
+Manifest files specify which agent to run:
+```yaml
+build_args:
+  AGENT_FILE: creator.py
+  PORT: 8001
+```
+
+### State Machine Flow
+The orchestrator coordinates the workflow through these states:
+- `CREATING` → `WAITING_FOR_CREATOR` → `REVIEWING` → `WAITING_FOR_CRITIC` → `FORMATTING` → `COMPLETED`
+
+### Inter-Agent Communication
+Agents communicate using AgentEx events:
+```python
+await adk.acp.send_event(
+    agent_name="ab090-creator-agent",
+    task_id=task_id,
+    content=TextContent(author="agent", content=json.dumps(request_data))
+)
+```
+
+## 📚 What You'll Learn
+
+This tutorial demonstrates:
+- **Multi-agent coordination** using state machines for complex workflows
+- **Type-safe communication** with Pydantic models for all request/response data
+- **Shared build configuration** for multiple agents in a single deployment
+- **AgentEx CLI usage** for development and deployment
+- **Inter-agent communication patterns** with proper error handling
+- **Scalable agent architecture** with clear separation of concerns