agent and tool documentatation update

Author: rawatSingh2022

.zshrc

@@ -0,0 +1,13 @@
+# Log fnm environment initialization
+fnm_env_with_logging() {
+  local log_file="$HOME/fnm_env.log"
+  echo "$(date): Running fnm env" >> "$log_file"
+  fnm env >> "$log_file" 2>&1
+  eval "$(fnm env)"
+}
+
+# Replace the original command with our logging function
+alias load_fnm='fnm_env_with_logging'
+
+# Uncomment to automatically run on shell startup
+# fnm_env_with_logging 

docs/agents/1_agentIntroduction.md

@@ -1,4 +1,4 @@
-# App Introduction
+# Agent Introduction
 
 Codebolt is an extremely powerful tool for AI Code Editing and AI Code Generation. Most of the power of Codebolt comes from the ability to use Agents. These agents are powerful tools that can be used to automate tasks, generate code, and even create entire applications. 
 
@@ -13,7 +13,7 @@ Codebolt Editor exposes all the functionality of the Editor in the format of API
 - [JavaScript](https://github.com/codebolt-ai/codebolt-js)
 - [Python](https://github.com/codebolt-ai/codebolt-python) -->
 
-## Codebolt Agents
+<!-- ## Codebolt Agents
 
 Codebolt has Multiple Types of Agents that can do various tasks.
 - [Universal Agents](./AgentTypes/UniversalAgents/universalAgents.md)
@@ -23,5 +23,5 @@ Codebolt has Multiple Types of Agents that can do various tasks.
 Universal Agents are called whenever the user sends any chat in the editor or asks to perform any action that can be resolved to any agent. You can learn more about them at [Universal Agents](./AgentTypes/UniversalAgents/universalAgents.md)
 
 ### Action Agents
-Action Agents are called whenever the user sends any chat in the editor or asks to perform any action that can be resolved to any agent. You can learn more about them at [Action Agents](./AgentTypes/actionAgents.md)
+Action Agents are called whenever the user sends any chat in the editor or asks to perform any action that can be resolved to any agent. You can learn more about them at [Action Agents](./AgentTypes/actionAgents.md) -->
 

docs/agents/2_firstExtension.md

@@ -1,45 +1,103 @@
-# Create your First Agent
+# Creating Your First Codebolt Agent
 
-## Create Codebolt Agent
-Follow the steps below to create a new agent using Codebolt.
+This guide will walk you through the process of creating your first custom Codebolt agent. Custom agents allow you to extend Codebolt's functionality with your own AI-powered tools.
 
-### Step 1: Select or Create a Project for Agent Creation
+## Prerequisites
 
-  ![create agent](../../static/img/create-agent.png)
+Before creating an agent, ensure you have:
+- Codebolt installed on your system
+- Node.js and npm (comes with the Codebolt installation)
+- Basic understanding of JavaScript/TypeScript
 
-- *After creating or opening a project, open a terminal to create an agent.*
+## Step-by-Step Guide
 
-### Step 2: Initialize the Agent Creation Process
+### 1. Initiating Agent Creation
 
-1. Open your terminal in the project directory
-2. Type the following command:
-   ```bash
-   npx codebolt-cli createagent demoAgent --quick
-   ```
- 
- ![open terminal](../../static/img/open-terminal.png)  
-3. Press Enter to execute the command
+To begin creating a new agent:
 
-   - `demoAgent`: This is the name of your agent (can be customized)
-   - `--quick`: This flag creates a basic agent template with default configurations
-### Step 3: Understanding the Agent Structure
+1. **Open Codebolt** and navigate to the chat interface.
+2. **Click on the "Create Agent" button** in the chat window. This option is typically found in the sidebar or main interface.
 
-After successful creation, your agent will have the following structure:
-```
-.codeboltAgents/
-└── demoAgent/
-    ├── codeboltagent.yaml       # Agent configuration file
-    ├── agent.yaml        # contains Prompt for agent
-    ├── task.yaml          # contains instruction for task
-```
+![In chat window click on create agent](../../static/img/create-agent.png)
 
-### Step 4: Verify Agent Creation
+### 2. Select Agent Type
 
-- *After successfully creating the agent, you can find your agent inside the .codeboltAgents folder.*
-  ![after success](../../static/img/after-creation.png)
+Next, you'll need to choose the type of agent you want to create:
+
+1. **Select "Custom Agent"** from the available options. This allows you to create an agent with your own defined functionality.
+
+![Select custom agent](../../static/img/custom-agent.png)
+
+### 3. Configure Your Agent
+
+Now it's time to set up the basic details for your agent:
+
+1. **Name Your Agent**: Enter a descriptive name that indicates the agent's purpose.
+2. **Add a Description**: Provide a brief description explaining what your agent does.
+
+![Name and description for agent](../../static/img/agent-name.png)
+
+### 4. Creation Confirmation
+
+After configuring your agent, the system will create the necessary files and structure:
+
+1. **Success Message**: You'll see a confirmation screen indicating that your agent has been successfully created.
+
+![Success confirmation screen](../../static/img/success.png)
+
+### 5. Explore Generated Code
+
+Codebolt will generate a starter code structure for your agent:
+
+1. **Code Structure**: Review the generated files. Typically, this includes:
+   - Main agent file with the core functionality
+   - Configuration files
+   - Dependencies and setup scripts
+
+![Generated code structure](../../static/img/code.png)
+
+### 6. Understanding the Agent Structure
+
+The generated code provides a foundation for your custom agent with the following key components:
+
+- **Main Agent File**: Contains the core logic for your agent
+- **Configuration Settings**: Defines how your agent interacts with Codebolt
+- **Tool Definitions**: Specifies the capabilities of your agent
+- **API Integration**: Enables your agent to communicate with external services
+
+### 7. Customizing Your Agent
+
+Now that your agent is created, you can start customizing it:
+
+1. **Modify the Agent Logic**: Update the generated code to implement your desired functionality
+2. **Add Custom Tools**: Define additional tools that your agent can use
+3. **Configure Settings**: Adjust the configuration to meet your specific requirements
+4. **Implement API Calls**: Add code to connect with external services or APIs if needed
+
+### 8. Testing Your Agent
+
+Before finalizing your agent, it's important to test it:
+
+1. **Run the Agent Locally**: Use the Codebolt testing environment to run your agent
+2. **Debug Issues**: Check for any errors or unexpected behavior
+3. **Refine Functionality**: Make adjustments based on testing results
+
+### 9. Deploying Your Agent
+
+Once your agent is working as expected, you can deploy it:
+
+1. **Build Your Agent**: Compile the final version of your agent
+2. **Register in Codebolt**: Add your agent to the Codebolt registry
+3. **Share with Others**: Optionally, share your agent with the Codebolt community
+
+## Next Steps
+
+After creating your first agent, consider:
+- Exploring advanced agent features
+- Creating agents for specific workflows or tasks
+- Learning about agent collaboration and chaining
+- Contributing to the Codebolt agent ecosystem
+
+By following this guide, you've successfully created your first Codebolt agent. As you become more familiar with the platform, you can develop increasingly sophisticated agents to enhance your productivity and workflow.
 
-### Step 5: Next Steps
 
-1. Open the `codeboltagent.yaml` file to customize your agent's settings
-2. Configure your agent's prompt in the `agent.yaml` file
-3. Define tasks and instructions in the `task.yaml` file

docs/agents/3_runExtension.md

@@ -1,37 +1,44 @@
-# Run Local Agent in Codebolt
+# Running Local Agents in Codebolt
 
 This guide will walk you through the process of running a local agent in Codebolt. Local agents allow you to use custom-built AI assistants directly within your development environment.
 
 ## Step-by-Step Guide
 
 ### 1. Access the Agent Menu
-- Open Codebolt and locate the "Agent" section in the main toolbar
-- Click on the "Agent: auto chat" header to open the agent selection menu
+- Open Codebolt and locate the "Agent" section in the chat window
+- Click on the "Agent" dropdown menu in the top bar
 
-![install agent](../../static/img/select-auto.png)
+![Access Agent Menu](../../static/img/browseAgent.png)
 
 ### 2. Navigate to Local Agents
-- In the popup window, you'll see three tabs:
-  - **Installed**: Shows currently active agents
-  - **MarketPlace**: Contains pre-built agents available for download
-  - **Local**: Displays agents stored on your machine
-- Click on the "Local" tab to view your available local agents
+- Click on "Browse agents" in the dropdown menu
+- Select "Local agent" from the available options
+
+![Select Local Agent](../../static/img/selectLocalAgent.png)
 
-![runlocal Agent](../../static/img/select-local-agent.png)
 ### 3. Select Your Agent
-- Browse through the list of local agents
+- Browse through the list of available local agents
 - Each agent will display its name and basic information
 - Click on the agent you want to activate
+- The selected agent will become your active assistant
+
+![Select From Local Agents](../../static/img/selectFormAllLocalAgent.png)
 
+### 4. Debug Your Agent
+- All debug logs will be displayed in the debug window
+- Use these logs to troubleshoot any issues with your agent's performance
+- Monitor the execution of your agent's actions in real-time
 
-![runlocal Agent](../../static/img/select-agent.png)
+![Debug Agent](../../static/img/DebugAgent.png)
 
+### 5. Interact with Your Agent
+- Once activated, you can communicate with your local agent directly in the chat window
+- Ask questions, request assistance with coding tasks, or use any specialized functionality your agent provides
+- Your agent will respond based on its programmed capabilities
 
-- After selection, the selected agent will be activated.
+## Additional Information
+- Local agents run directly on your machine, ensuring your code and queries remain private
+- You can switch between different agents at any time by repeating the selection process
+- For information on creating your own custom agent, refer to our [Agent Development Guide](./agentDevelopment.md)
 
 
-![runlocal Agent](../../static/img/after-selection.png)
-## Troubleshooting Tips
-- If you don't see your local agent:
-  - Verify the agent files are in the correct directory
-  - Restart Codebolt and try again

docs/agents/4_runAgent.md

@@ -11,7 +11,7 @@ To use or run an agent, you’ll need to follow these steps:
 
 * **Switch Agents as Needed**: If you need a different set of functions, return to your agent list to deactivate the current agent and select another. For instance, if you’re done with development and need to deploy, switch from CodeBolt Dev to CodeBolt Web Deployment for automated deployment support.
 
-![selectAgent](../../static/img/selectAgent.png)
+![selectAgent](../../static/img/browseAgent.png)
 
 **To use agent in chat, you can follow these steps:**
 
@@ -23,4 +23,6 @@ In the chat interface, click on the `# symbol`. This will open list of all avail
 
 * **Switch Agents if Needed**: If you need help with a different type of task, click # again to open the agent list, select another agent, and start your next interaction.
 
-![runAgent2](../../static/img/runAgent2.png)
+![runAgent2](../../static/img/selectMarketplaceAgent.png)
+
+![runAgent2](../../static/img/SelectFormAgent.png)

docs/tools/1_tools.md

@@ -0,0 +1,11 @@
+# Tools
+
+Codebolt supports writing your own custom Model Control Protocol (MCP) tools. These custom tools allow you to extend Codebolt's functionality by creating specialized actions that AI agents can perform. You can define your own tool interfaces, implement their functionality, and make them available for agents to use when solving tasks.
+
+Custom MCP tools are powerful for:
+- Integrating with external services and APIs
+- Creating domain-specific functionality for your projects
+- Automating repetitive tasks with specialized tools
+- Building workflow-specific capabilities
+
+

docs/tools/2_create_tool.md

@@ -0,0 +1,57 @@
+# Creating Your First Tool
+
+This guide walks you through the process of creating and customizing your first Codebolt tool.
+
+## Tool Creation
+
+You can create a new tool using either the Codebolt UI or the CLI:
+
+### Option 1: Using the Codebolt UI
+
+Navigate to the Codebolt dashboard and use the tool creation interface:
+
+![Codebolt tool creation interface](../../static/img/mcpcreate.png)
+
+Enter a descriptive name for your tool:
+
+![Naming your Codebolt tool](../../static/img/nameyourmcp.png)
+
+### Option 2: Using the CLI
+
+Alternatively, use the Codebolt CLI by running:
+
+```bash
+npx codebolt-cli createtool <toolName>
+```
+
+Replace `<toolName>` with a descriptive name for your tool.
+
+![CLI tool creation example](../../static/img/createTool.png)
+
+## Tool Configuration
+
+After creation, your tool will be available in the `.codeboltAgent/tool` directory within your project structure. This directory contains all configuration and implementation files needed for your tool.
+
+![Tool directory structure](../../static/img/modifyyourtool.png)
+
+## Customization Options
+
+You can customize various aspects of your tool:
+
+- Input parameters and validation
+- Processing logic and business rules
+- Output formatting and response handling
+- Integration with external services
+
+For detailed information on available configuration options, refer to the [Tool Configuration Reference](../reference/tool-configuration.md).
+
+## Testing and Deployment
+
+Once configured, your tool will appear in the Tools section of the Codebolt chat interface. You can:
+
+1. Test the tool's functionality with sample inputs
+2. Debug any issues in real-time
+3. Refine the tool's behavior as needed
+4. Deploy for production use
+
+This iterative development process ensures your tool performs as expected in your application ecosystem.

docs/tools/3_testlocalmcp.md

@@ -0,0 +1,40 @@
+# Testing Your Local MCP
+
+This guide will walk you through the process of testing a locally deployed MCP (Modular Capability Processor).
+
+
+
+## Testing Options
+
+You can test your local MCP using either the graphical interface or the command-line interface.
+
+### Option 1: Using the GUI
+
+#### 1. Open the MCP Panel
+
+Access the MCP panel by clicking the MCP icon in the Codebolt interface.
+
+![Open MCP Panel](../../static/img/openMcp.png)
+
+#### 2. Select Local MCP
+
+From the dropdown menu, select "Local MCP" to connect to your locally running instance.
+
+![Select Local MCP](../../static/img/SelectMcp.png)
+
+### Option 2: Using the CLI
+
+You can also test specific tools directly from the command line:
+
+```bash
+codebolt-cli testtool <toolName>
+```
+
+Replace `<toolName>` with the name of the specific tool you want to test.
+
+## Verifying Connection
+
+Once connected, check for status indicators confirming a successful connection to your local MCP instance.
+
+
+

docusaurus.config.ts

@@ -74,6 +74,12 @@ const config: Config = {
           position: 'left',
           label: 'Agents',
         },
+        {
+          type: 'docSidebar',
+          sidebarId: 'toolSidebar',
+          position: 'left',
+          label: 'Tools',
+        },
         {
           type: 'docSidebar',
           sidebarId: 'appSidebar',

fnm_env_output.log

@@ -0,0 +1,10 @@
+export PATH="/Users/ravirawat/.local/state/fnm_multishells/35935_1746877594537/bin":$PATH
+export FNM_MULTISHELL_PATH="/Users/ravirawat/.local/state/fnm_multishells/35935_1746877594537"
+export FNM_VERSION_FILE_STRATEGY="local"
+export FNM_DIR="/Users/ravirawat/.local/share/fnm"
+export FNM_LOGLEVEL="info"
+export FNM_NODE_DIST_MIRROR="https://nodejs.org/dist"
+export FNM_COREPACK_ENABLED="false"
+export FNM_RESOLVE_ENGINES="true"
+export FNM_ARCH="arm64"
+rehash