massgen 0.1.4__py3-none-any.whl → 0.1.6__py3-none-any.whl

massgen/configs/memory/single_agent_compression_test.yaml

@@ -0,0 +1,64 @@
+# Single Agent Compression Test
+#
+# Simple config to test per-agent memory and compression with ONE agent.
+# Set very low thresholds to trigger compression quickly.
+#
+# Run with:
+#   uv run massgen --config massgen/configs/memory/single_agent_compression_test.yaml "Analyze the MassGen codebase comprehensively. Create an architecture document that explains: (1) Core components and their responsibilities, (2) How different modules interact, (3) Key design patterns used, (4) Main entry points and request flows. Read > 30 files to build a complete understanding."
+
+agents:
+  - id: "test_agent"
+    backend:
+      type: "openai"
+      model: "gpt-5-nano"
+      cwd: "workspace1"
+
+# Global memory configuration
+memory:
+  enabled: true
+
+  conversation_memory:
+    enabled: true
+
+  persistent_memory:
+    enabled: true
+
+    # LLM for memory operations
+    llm:
+      provider: "openai"
+      model: "gpt-4.1-nano-2025-04-14"
+
+    # Embedding for vector search
+    embedding:
+      provider: "openai"
+      model: "text-embedding-3-small"
+
+    # Qdrant configuration (use server mode)
+    qdrant:
+      mode: "server"
+      host: "localhost"
+      port: 6333
+
+  compression:
+    trigger_threshold: 0.02   # Compress at 2% (very low for quick testing)
+    target_ratio: 0.01      # Keep 1% after compression
+
+  # Memory retrieval configuration
+  retrieval:
+    limit: 5              # Number of memory facts to retrieve from mem0 (default: 5)
+    exclude_recent: true  # Only retrieve after compression to avoid duplicates (default: true)
+                          # When true: retrieval skipped until compression occurs
+                          # When false: always retrieve (may duplicate recent context)
+
+# Orchestrator config
+orchestrator:
+  snapshot_storage: "test_compression_sessions"
+  agent_temporary_workspace: "test_compression_workspace"
+  context_paths:
+    - path: "massgen"
+      permission: "read"
+
+# UI
+ui:
+  display_type: "rich_terminal"
+  logging_enabled: true

massgen/configs/tools/custom_tools/claude_code_custom_tool_with_mcp_example.yaml

@@ -1,6 +1,7 @@
 # MassGen Configuration: Claude Code with MCP Integration
 # Usage:
 #   massgen --config @examples/tools/custom_tools/claude_code_custom_tool_with_mcp_example "whats the sum of 123 and 456? and whats the weather of Tokyo?"
+#   uv run python -m massgen.cli --config massgen/configs/tools/custom_tools/claude_code_custom_tool_with_mcp_example.yaml "whats the sum of 123 and 456? and whats the weather of Tokyo?"
 orchestrator:
   snapshot_storage: "claude_code_snapshots"
   agent_temporary_workspace: "claude_code_temp"

massgen/configs/tools/custom_tools/claude_custom_tool_example_no_path.yaml

@@ -18,5 +18,5 @@ agents:
       When users ask about adding two numbers together, use the two_num_tool to calculate the sum and provide
       a clear, accurate result.
 ui:
-  display_type: "simple"
+  display_type: "rich_terminal"
   logging_enabled: true

massgen/configs/tools/custom_tools/claude_custom_tool_with_mcp_example.yaml

@@ -1,6 +1,7 @@
 # MassGen Configuration: Claude with MCP Integration
 # Usage:
 #   massgen --config @examples/tools/custom_tools/claude_custom_tool_with_mcp_example "whats the sum of 123 and 456? and whats the weather of Tokyo?"
+#   uv run python -m massgen.cli --config massgen/configs/tools/custom_tools/claude_custom_tool_with_mcp_example.yaml "whats the sum of 123 and 456? and whats the weather of Tokyo?"
 agents:
   - id: "claude_custom_tool" # Anthropic Claude
     backend:

massgen/configs/tools/custom_tools/computer_use_browser_example.yaml

@@ -51,6 +51,6 @@ agents:
       - Be specific about what you're looking for on the page
 
 ui:
-  display_type: "detailed"
+  display_type: "rich_terminal"
   logging_enabled: true
   show_screenshots: true

massgen/configs/tools/custom_tools/computer_use_docker_example.yaml

@@ -59,7 +59,7 @@ agents:
       - Consider screen resolution (1280x800)
 
 ui:
-  display_type: "detailed"
+  display_type: "rich_terminal"
   logging_enabled: true
   show_screenshots: true
   show_reasoning: true

massgen/configs/tools/custom_tools/gemini_custom_tool_with_mcp_example.yaml

@@ -1,6 +1,7 @@
 # MassGen Configuration: Gemini with MCP Integration
 # Usage:
 #   massgen --config @examples/tools/custom_tools/gemini_custom_tool_with_mcp_example "whats the sum of 123 and 456? and whats the weather of Tokyo?"
+#   uv run python -m massgen.cli --config massgen/configs/tools/custom_tools/gemini_custom_tool_with_mcp_example.yaml "whats the sum of 123 and 456? and whats the weather of Tokyo?"
 agents:
   - id: "gemini2.5flash_custom_tool" # Google Gemini 2.5 Flash
     backend:

massgen/configs/tools/custom_tools/gpt5_nano_custom_tool_with_mcp_example.yaml

@@ -1,6 +1,7 @@
 # MassGen Configuration: GPT-5 Nano with MCP Integration
 # Usage:
 #   massgen --config @examples/tools/custom_tools/gpt5_nano_custom_tool_with_mcp_example "whats the sum of 123 and 456? and whats the weather of Tokyo?"
+#   uv run python -m massgen.cli --config massgen/configs/tools/custom_tools/gpt5_nano_custom_tool_with_mcp_example.yaml "whats the sum of 123 and 456? and whats the weather of Tokyo?"
 agents:
   - id: "gpt5_nano_custom_tool" # OpenAI GPT-5 Nano
     backend:

massgen/configs/tools/custom_tools/gpt_oss_custom_tool_with_mcp_example.yaml

@@ -1,6 +1,7 @@
 # MassGen Configuration: Gemini with MCP Integration
 # Usage:
 #   massgen --config @examples/tools/custom_tools/gpt_oss_custom_tool_with_mcp_example "whats the sum of 123 and 456? and whats the weather of Tokyo?"
+#   uv run python -m massgen.cli --config massgen/configs/tools/custom_tools/gpt_oss_custom_tool_with_mcp_example.yaml "whats the sum of 123 and 456? and whats the weather of Tokyo?"
 agents:
   - id: "gpt-oss-custom_tool" # Cerebras AI GPT-OSS
     backend:

massgen/configs/tools/custom_tools/grok3_mini_custom_tool_with_mcp_example.yaml

@@ -1,6 +1,7 @@
 # MassGen Configuration: Grok-3 Mini with MCP Integration
 # Usage:
 #   massgen --config @examples/tools/custom_tools/grok3_mini_custom_tool_with_mcp_example "whats the sum of 123 and 456? and whats the weather of Tokyo?"
+#   uv run python -m massgen.cli --config massgen/configs/tools/custom_tools/grok3_mini_custom_tool_with_mcp_example.yaml "whats the sum of 123 and 456? and whats the weather of Tokyo?"
 agents:
   - id: "grok3_mini_custom_tool"
     backend:

massgen/configs/tools/custom_tools/interop/ag2_and_langgraph_lesson_planner.yaml

@@ -0,0 +1,65 @@
+# MassGen Configuration: Lesson Planner Comparison
+# This demonstrates interoperability by comparing AG2 and LangGraph lesson planner implementations
+#
+# Usage:
+#   massgen --config @examples/tools/custom_tools/interop/ag2_and_langgraph_lesson_planner.yaml "Create a lesson plan for photosynthesis"
+#   uv run python -m massgen.cli --config @examples/tools/custom_tools/interop/ag2_and_langgraph_lesson_planner.yaml "Create a lesson plan for photosynthesis"
+#
+# Description:
+#   This configuration provides access to both AG2 (AutoGen) and LangGraph lesson planning tools,
+#   allowing you to compare different multi-agent framework approaches within MassGen.
+#
+#   AG2 Version:
+#   - Uses AutoGen's nested chat architecture
+#   - Multiple agents in group chat collaboration
+#   - Demonstrates AutoGen → MassGen interoperability
+#
+#   LangGraph Version:
+#   - Uses LangGraph's state graph architecture
+#   - Sequential workflow with state management
+#   - Demonstrates LangGraph → MassGen interoperability
+#
+# Requirements:
+#   - pip install pyautogen langgraph langchain-openai langchain-core
+#   - OPENAI_API_KEY environment variable must be set
+
+agents:
+  - id: "ag2_planner" # AG2 (AutoGen) implementation
+    backend:
+      type: "openai"
+      model: "gpt-4o"
+      custom_tools:
+        - name: ["ag2_lesson_planner"]
+          category: "education"
+          path: "massgen/tool/_extraframework_agents/ag2_lesson_planner_tool.py"
+          function: ["ag2_lesson_planner"]
+    system_message: |
+      You are an AI assistant with access to the AG2 (AutoGen) lesson planning tool.
+
+      The ag2_lesson_planner tool creates comprehensive fourth-grade lesson plans using AutoGen's
+      nested chat architecture with multiple collaborative agents.
+
+      When users ask for a lesson plan, use the ag2_lesson_planner tool. After receiving the result,
+      present it clearly and mention that this was generated using the AG2 (AutoGen) framework.
+
+  - id: "langgraph_planner" # LangGraph implementation
+    backend:
+      type: "openai"
+      model: "gpt-4o"
+      custom_tools:
+        - name: ["langgraph_lesson_planner"]
+          category: "education"
+          path: "massgen/tool/_extraframework_agents/langgraph_lesson_planner_tool.py"
+          function: ["langgraph_lesson_planner"]
+    system_message: |
+      You are an AI assistant with access to the LangGraph lesson planning tool.
+
+      The langgraph_lesson_planner tool creates comprehensive fourth-grade lesson plans using
+      LangGraph's state graph architecture with sequential workflow nodes.
+
+      When users ask for a lesson plan, use the langgraph_lesson_planner tool. After receiving the result,
+      present it clearly and mention that this was generated using the LangGraph framework.
+
+ui:
+  display_type: "rich_terminal"
+  logging_enabled: true

massgen/configs/tools/custom_tools/interop/ag2_and_openai_assistant_lesson_planner.yaml

@@ -0,0 +1,65 @@
+# MassGen Configuration: Lesson Planner Comparison
+# This demonstrates interoperability by comparing AG2 and OpenAI Assistant lesson planner implementations
+#
+# Usage:
+#   massgen --config @examples/tools/custom_tools/interop/ag2_and_openai_assistant_lesson_planner.yaml "Create a lesson plan for photosynthesis"
+#   uv run python -m massgen.cli --config @examples/tools/custom_tools/interop/ag2_and_openai_assistant_lesson_planner.yaml "Create a lesson plan for photosynthesis"
+#
+# Description:
+#   This configuration provides access to both AG2 (AutoGen) and OpenAI Assistant lesson planning tools,
+#   allowing you to compare different multi-agent framework approaches within MassGen.
+#
+#   AG2 Version:
+#   - Uses AutoGen's nested chat architecture
+#   - Multiple agents in group chat collaboration
+#   - Demonstrates AutoGen → MassGen interoperability
+#
+#   OpenAI Assistant Version:
+#   - Uses OpenAI's Chat Completions API with streaming
+#   - Multi-agent collaboration pattern with sequential workflow
+#   - Demonstrates OpenAI Assistant → MassGen interoperability
+#
+# Requirements:
+#   - pip install pyautogen openai
+#   - OPENAI_API_KEY environment variable must be set
+
+agents:
+  - id: "ag2_planner" # AG2 (AutoGen) implementation
+    backend:
+      type: "openai"
+      model: "gpt-4o"
+      custom_tools:
+        - name: ["ag2_lesson_planner"]
+          category: "education"
+          path: "massgen/tool/_extraframework_agents/ag2_lesson_planner_tool.py"
+          function: ["ag2_lesson_planner"]
+    system_message: |
+      You are an AI assistant with access to the AG2 (AutoGen) lesson planning tool.
+
+      The ag2_lesson_planner tool creates comprehensive fourth-grade lesson plans using AutoGen's
+      nested chat architecture with multiple collaborative agents.
+
+      When users ask for a lesson plan, use the ag2_lesson_planner tool. After receiving the result,
+      present it clearly and mention that this was generated using the AG2 (AutoGen) framework.
+
+  - id: "openai_assistant_planner" # OpenAI Assistant implementation
+    backend:
+      type: "openai"
+      model: "gpt-4o"
+      custom_tools:
+        - name: ["openai_assistant_lesson_planner"]
+          category: "education"
+          path: "massgen/tool/_extraframework_agents/openai_assistant_lesson_planner_tool.py"
+          function: ["openai_assistant_lesson_planner"]
+    system_message: |
+      You are an AI assistant with access to the OpenAI Assistant lesson planning tool.
+
+      The openai_assistant_lesson_planner tool creates comprehensive fourth-grade lesson plans using
+      OpenAI's Chat Completions API with multi-agent collaboration pattern and streaming support.
+
+      When users ask for a lesson plan, use the openai_assistant_lesson_planner tool. After receiving the result,
+      present it clearly and mention that this was generated using the OpenAI Assistant framework.
+
+ui:
+  display_type: "rich_terminal"
+  logging_enabled: true

massgen/configs/tools/custom_tools/interop/ag2_lesson_planner_example.yaml

@@ -0,0 +1,48 @@
+# MassGen Configuration: AG2 (AutoGen) Lesson Planner Tool
+# This demonstrates interoperability by using AutoGen's nested chat as a custom tool
+#
+# Usage:
+#   massgen --config @examples/tools/custom_tools/interop/ag2_lesson_planner_example.yaml "Create a lesson plan for photosynthesis"
+#   uv run python -m massgen.cli --config @examples/tools/custom_tools/interop/ag2_lesson_planner_example.yaml "Create a lesson plan for photosynthesis"
+#
+# Description:
+#   This configuration uses the AG2 (AutoGen) nested chat architecture to create comprehensive
+#   lesson plans. The tool uses multiple AutoGen agents working together:
+#   - Curriculum Agent: Determines standards and objectives
+#   - Lesson Planner Agent: Creates detailed lesson plans
+#   - Lesson Reviewer Agent: Reviews and provides feedback
+#   - Lesson Formatter Agent: Formats the final output
+#
+# Requirements:
+#   - pip install pyautogen
+#   - OPENAI_API_KEY environment variable must be set
+
+agents:
+  - id: "ag2_lesson_planner" # OpenAI GPT-4o with AG2 custom tool
+    backend:
+      type: "openai"
+      model: "gpt-4o"
+      custom_tools:
+        - name: ["ag2_lesson_planner"]
+          category: "education"
+          path: "massgen/tool/_extraframework_agents/ag2_lesson_planner_tool.py"
+          function: ["ag2_lesson_planner"]
+    system_message: |
+      You are an AI assistant with access to an advanced lesson planning tool powered by AutoGen.
+
+      The ag2_lesson_planner tool is available as a custom tool that creates comprehensive fourth-grade
+      lesson plans using a multi-agent collaborative approach. This tool demonstrates interoperability
+      between MassGen and AutoGen frameworks.
+
+      When users ask for a lesson plan on any topic, use the ag2_lesson_planner tool. The tool will coordinate multiple specialized agents to:
+      1. Identify relevant curriculum standards and learning objectives
+      2. Design engaging lesson activities with proper timing
+      3. Review the plan for age-appropriateness and effectiveness
+      4. Format the final lesson plan in a structured format
+
+      After receiving the lesson plan, present it clearly to the user and offer to make any adjustments
+      they'd like.
+
+ui:
+  display_type: "rich_terminal"
+  logging_enabled: true

massgen/configs/tools/custom_tools/interop/agentscope_lesson_planner_example.yaml

@@ -0,0 +1,48 @@
+# MassGen Configuration: AgentScope Lesson Planner Tool
+# This demonstrates interoperability by using AgentScope's multi-agent framework as a custom tool
+#
+# Usage:
+#   massgen --config @examples/tools/custom_tools/interop/agentscope_lesson_planner_example.yaml "Create a lesson plan for photosynthesis"
+#   uv run python -m massgen.cli --config @examples/tools/custom_tools/interop/agentscope_lesson_planner_example.yaml "Create a lesson plan for photosynthesis"
+#
+# Description:
+#   This configuration uses AgentScope to create comprehensive lesson plans through a multi-agent
+#   sequential pipeline. The tool uses multiple specialized AgentScope DialogAgents:
+#   - Curriculum Standards Expert: Determines standards and objectives
+#   - Lesson Planning Specialist: Creates detailed lesson plans
+#   - Lesson Plan Reviewer: Reviews and improves the plan
+#   - Lesson Plan Formatter: Formats the final output
+#
+# Requirements:
+#   - pip install agentscope
+#   - OPENAI_API_KEY environment variable must be set
+
+agents:
+  - id: "agentscope_lesson_planner" # OpenAI GPT-4o with AgentScope custom tool
+    backend:
+      type: "openai"
+      model: "gpt-4o"
+      custom_tools:
+        - name: ["agentscope_lesson_planner"]
+          category: "education"
+          path: "massgen/tool/_extraframework_agents/agentscope_lesson_planner_tool.py"
+          function: ["agentscope_lesson_planner"]
+    system_message: |
+      You are an AI assistant with access to an advanced lesson planning tool powered by AgentScope.
+
+      The agentscope_lesson_planner tool is available as a custom tool that creates comprehensive fourth-grade
+      lesson plans using AgentScope's multi-agent framework. This tool demonstrates interoperability
+      between MassGen and AgentScope.
+
+      When users ask for a lesson plan on any topic, use the agentscope_lesson_planner tool. The tool will coordinate multiple specialized agents in a sequential pipeline to:
+      1. Identify relevant curriculum standards and learning objectives
+      2. Design engaging lesson activities with proper timing
+      3. Review the plan for age-appropriateness and effectiveness
+      4. Format the final lesson plan in a structured format
+
+      After receiving the lesson plan, present it clearly to the user and offer to make any adjustments
+      they'd like.
+
+ui:
+  display_type: "rich_terminal"
+  logging_enabled: true

massgen/configs/tools/custom_tools/interop/langgraph_lesson_planner_example.yaml

@@ -0,0 +1,49 @@
+# MassGen Configuration: LangGraph Lesson Planner Tool
+# This demonstrates interoperability by using LangGraph's state graph framework as a custom tool
+#
+# Usage:
+#   massgen --config @examples/tools/custom_tools/interop/langgraph_lesson_planner_example.yaml "Create a lesson plan for photosynthesis"
+#   uv run python -m massgen.cli --config @examples/tools/custom_tools/interop/langgraph_lesson_planner_example.yaml "Create a lesson plan for photosynthesis"
+#
+# Description:
+#   This configuration uses LangGraph to create comprehensive lesson plans through a state graph
+#   workflow. The tool orchestrates multiple specialized nodes in a sequential graph:
+#   - Curriculum Node: Determines standards and objectives
+#   - Planner Node: Creates detailed lesson plans
+#   - Reviewer Node: Reviews and improves the plan
+#   - Formatter Node: Formats the final output
+#
+# Requirements:
+#   - pip install langgraph langchain-openai langchain-core
+#   - OPENAI_API_KEY environment variable must be set
+
+agents:
+  - id: "langgraph_lesson_planner" # OpenAI GPT-4o with LangGraph custom tool
+    backend:
+      type: "openai"
+      model: "gpt-4o"
+      custom_tools:
+        - name: ["langgraph_lesson_planner"]
+          category: "education"
+          path: "massgen/tool/_extraframework_agents/langgraph_lesson_planner_tool.py"
+          function: ["langgraph_lesson_planner"]
+    system_message: |
+      You are an AI assistant with access to an advanced lesson planning tool powered by LangGraph.
+
+      The langgraph_lesson_planner tool is available as a custom tool that creates comprehensive fourth-grade
+      lesson plans using LangGraph's state graph architecture. This tool demonstrates interoperability
+      between MassGen and LangGraph/LangChain frameworks.
+
+      When users ask for a lesson plan on any topic, use the langgraph_lesson_planner tool. The tool will execute
+      a state graph workflow with multiple nodes to:
+      1. Identify relevant curriculum standards and learning objectives
+      2. Design engaging lesson activities with proper timing
+      3. Review the plan for age-appropriateness and effectiveness
+      4. Format the final lesson plan in a structured format
+
+      After receiving the lesson plan, present it clearly to the user and offer to make any adjustments
+      they'd like.
+
+ui:
+  display_type: "rich_terminal"
+  logging_enabled: true

massgen/configs/tools/custom_tools/interop/openai_assistant_lesson_planner_example.yaml

@@ -0,0 +1,50 @@
+# MassGen Configuration: OpenAI Lesson Planner Tool
+# This demonstrates interoperability by using OpenAI's Chat Completions API as a custom tool
+#
+# Usage:
+#   massgen --config @examples/tools/custom_tools/interop/openai_assistant_lesson_planner_example.yaml "Create a lesson plan for photosynthesis"
+#   uv run python -m massgen.cli --config @examples/tools/custom_tools/interop/openai_assistant_lesson_planner_example.yaml "Create a lesson plan for photosynthesis"
+#
+# Description:
+#   This configuration uses OpenAI's Chat Completions API with multi-agent collaboration to create comprehensive lesson plans.
+#   The tool uses multiple specialized agents (similar to AG2) that work sequentially:
+#   1. Curriculum Agent: Identifies relevant curriculum standards and learning objectives
+#   2. Lesson Planner Agent: Designs engaging lesson activities with proper timing
+#   3. Lesson Reviewer Agent: Reviews the plan for age-appropriateness and effectiveness
+#   4. Formatter Agent: Formats the final lesson plan in a structured format
+#   This demonstrates interoperability between MassGen and OpenAI's API using a role-based collaboration pattern.
+#
+# Requirements:
+#   - pip install openai
+#   - OPENAI_API_KEY environment variable must be set
+
+agents:
+  - id: "openai_assistant_lesson_planner" # OpenAI GPT-4o with OpenAI Assistant custom tool
+    backend:
+      type: "openai"
+      model: "gpt-4o"
+      custom_tools:
+        - name: ["openai_assistant_lesson_planner"]
+          category: "education"
+          path: "massgen/tool/_extraframework_agents/openai_assistant_lesson_planner_tool.py"
+          function: ["openai_assistant_lesson_planner"]
+    system_message: |
+      You are an AI assistant with access to an advanced lesson planning tool powered by OpenAI's Chat Completions API.
+
+      The openai_assistant_lesson_planner tool is available as a custom tool that creates comprehensive fourth-grade
+      lesson plans using OpenAI's Chat Completions API with multi-agent collaboration (similar to AG2/AutoGen pattern).
+      This tool demonstrates interoperability between MassGen and OpenAI's framework.
+
+      When users ask for a lesson plan on any topic, use the openai_assistant_lesson_planner tool. The tool uses
+      multiple specialized agents that collaborate sequentially:
+      1. Curriculum Agent: Identifies relevant curriculum standards and learning objectives
+      2. Lesson Planner Agent: Designs engaging lesson activities with proper timing
+      3. Lesson Reviewer Agent: Reviews the plan for age-appropriateness and effectiveness
+      4. Formatter Agent: Formats the final lesson plan in a structured format
+
+      After receiving the lesson plan, present it clearly to the user and offer to make any adjustments
+      they'd like.
+
+ui:
+  display_type: "rich_terminal"
+  logging_enabled: true

massgen/configs/tools/custom_tools/interop/smolagent_lesson_planner_example.yaml

@@ -0,0 +1,49 @@
+# MassGen Configuration: SmolAgent Lesson Planner Tool
+# This demonstrates interoperability by using HuggingFace's SmolAgent as a custom tool
+#
+# Usage:
+#   massgen --config @examples/tools/custom_tools/interop/smolagent_lesson_planner_example.yaml "Create a lesson plan for photosynthesis"
+#   uv run python -m massgen.cli --config @examples/tools/custom_tools/interop/smolagent_lesson_planner_example.yaml "Create a lesson plan for photosynthesis"
+#
+# Description:
+#   This configuration uses SmolAgent (HuggingFace's lightweight agent framework) to create comprehensive
+#   lesson plans. The tool uses SmolAgent's CodeAgent with custom tools:
+#   - get_curriculum_standards: Determines standards and objectives
+#   - create_lesson_plan: Creates detailed lesson plans
+#   - review_lesson_plan: Reviews and provides feedback
+#   - format_lesson_plan: Formats the final output
+#
+# Requirements:
+#   - pip install smolagents
+#   - OPENAI_API_KEY environment variable must be set
+
+agents:
+  - id: "smolagent_lesson_planner" # OpenAI GPT-4o with SmolAgent custom tool
+    backend:
+      type: "openai"
+      model: "gpt-4o"
+      custom_tools:
+        - name: ["smolagent_lesson_planner"]
+          category: "education"
+          path: "massgen/tool/_extraframework_agents/smolagent_lesson_planner_tool.py"
+          function: ["smolagent_lesson_planner"]
+    system_message: |
+      You are an AI assistant with access to an advanced lesson planning tool powered by SmolAgent.
+
+      The smolagent_lesson_planner tool is available as a custom tool that creates comprehensive fourth-grade
+      lesson plans using HuggingFace's SmolAgent framework. This tool demonstrates interoperability
+      between MassGen and SmolAgent.
+
+      When users ask for a lesson plan on any topic, use the smolagent_lesson_planner tool. The tool will use
+      multiple custom tools to:
+      1. Identify relevant curriculum standards and learning objectives
+      2. Design engaging lesson activities with proper timing
+      3. Review the plan for age-appropriateness and effectiveness
+      4. Format the final lesson plan in a structured format
+
+      After receiving the lesson plan, present it clearly to the user and offer to make any adjustments
+      they'd like.
+
+ui:
+  display_type: "rich_terminal"
+  logging_enabled: true

massgen/configs/tools/custom_tools/qwen_api_custom_tool_with_mcp_example.yaml

@@ -1,6 +1,7 @@
 # MassGen Configuration: Qwen API with MCP Integration
 # Usage:
 #   massgen --config @examples/tools/custom_tools/qwen_api_custom_tool_with_mcp_example "whats the sum of 123 and 456? and whats the weather of Tokyo?"
+#   uv run python -m massgen.cli --config massgen/configs/tools/custom_tools/qwen_api_custom_tool_with_mcp_example.yaml "whats the sum of 123 and 456? and whats the weather of Tokyo?"
 agents:
   - id: "qwen-mcp_weather" # Cerebras AI
     backend:

massgen/configs/tools/custom_tools/two_models_with_tools_example.yaml

@@ -0,0 +1,44 @@
+# MassGen Configuration: Two Models with Custom Tools
+# Usage:
+#   massgen --config massgen/configs/tools/custom_tools/two_models_with_tools_example.yaml "whats the sum of 123 and 456, and also 789 and 012?"
+#   uv run python -m massgen.cli --config massgen/configs/tools/custom_tools/two_models_with_tools_example.yaml "whats the sum of 123 and 456, and also 789 and 012?"
+agents:
+  - id: "claude_agent_with_tools" # Anthropic Claude Agent
+    backend:
+      type: "claude"
+      model: "claude-sonnet-4-20250514"
+      custom_tools:
+        - name: "two_num_tool"
+          category: "math"
+          path: "massgen/tool/_basic/_two_num_tool.py"
+          function: "two_num_tool"
+    system_message: |
+      You are Claude, an AI assistant with access to custom math calculation tools.
+
+      The two_num_tool is available as a custom tool that can add two numbers together.
+      Do not output tool call syntax or function declarations. Focus on answering the user's question clearly.
+
+      When users ask about adding two numbers together, use the two_num_tool to calculate the sum and provide
+      a clear, accurate result.
+
+  - id: "gemini_agent_with_tools" # Google Gemini Agent
+    backend:
+      type: "gemini"
+      model: "gemini-2.5-flash"
+      custom_tools:
+        - name: ["two_num_tool"]
+          category: "math"
+          path: "massgen/tool/_basic/_two_num_tool.py"
+          function: ["two_num_tool"]
+    system_message: |
+      You are Gemini, an AI assistant with access to custom math calculation tools.
+
+      The two_num_tool is available as a custom tool that can add two numbers together.
+      Do not output tool call syntax or function declarations. Focus on answering the user's question clearly.
+
+      When users ask about adding two numbers together, use the two_num_tool to calculate the sum and provide
+      a clear, accurate result.
+
+ui:
+  display_type: "rich_terminal"
+  logging_enabled: true