Update CLI documentation with new features since v1.2.0

openhands/usage/run-openhands/cli-mode.mdx

@@ -44,8 +44,8 @@ configuration format has changed.
     ```
   </Tab>
   <Tab title="Executable Binary">
-      1. Download the executable binary from
-     [OpenHands release page](https://github.com/OpenHands/OpenHands/releases/tag/1.0.1-cli) and rename it to
+      1. Download the executable binary from the
+     [OpenHands CLI release page](https://github.com/OpenHands/OpenHands-CLI/releases/latest) and rename it to
      `openhands` for simplicity.
 
       2. Make it executable:
@@ -117,12 +117,71 @@ The conversation history will be saved in `~/.openhands/conversations`.
 CLI mode enables real-time interaction with OpenHands through the terminal. You can type natural language tasks,
 use interactive commands, and receive instant feedback—all inside your terminal.
 
-### CLI Commands and Settings
+## CLI Options and Arguments
+
+The OpenHands CLI supports various command-line options to customize your experience:
+
+### Basic Options
+
+- `--version` or `-v`: Display the CLI version
+- `--task "your task"`: Start the CLI with an initial task message
+- `--file path/to/file`: Start the CLI with a specific file context
+- `--headless`: Run in headless mode with auto-exit functionality
+- `--exit-without-confirmation`: Exit without confirmation prompts
+
+### Experimental UI
+
+<Note>
+The experimental UI is available starting from CLI version 1.5.0 and provides an enhanced terminal experience.
+</Note>
+
+Enable the experimental UI with the `--exp` flag:
+
+```bash
+openhands --exp
+```
+
+**Experimental UI Features:**
+- **Enhanced Visual Experience**: Consistent dark mode CLI UI that overrides light-themed terminals
+- **Loading Indicators**: Spinner animations for long-running commands
+- **Smart Queuing**: Queue prompts while the agent is running to avoid interruptions
+- **File Search**: Use the `@` macro to search and reference files in your project
+- **Collapsible Messages**: Reduced verbosity with optionally collapsible CLI messages
+- **Multi-line Support**: Improved multi-line input handling with automatic paste detection
+
+### Approval and Confirmation Modes
+
+Control how the CLI handles agent actions:
+
+- `--always-approve`: Automatically approve all agent actions without prompting
+- `--llm-approve`: Use the LLM to automatically approve actions based on context
+
+
+
+## Skills and Capabilities
+
+The CLI automatically loads and provides access to:
+
+- **Project Skills**: Custom skills defined in your project's `.openhands/skills/` directory
+- **User Skills**: Personal skills from `~/.openhands/skills/`
+- **Public Skills**: Community-contributed skills from the OpenHands skills repository
+- **Legacy Microagents**: Backward compatibility with existing microagent configurations
+
+## Version Management
+
+The CLI includes automatic version checking and will prompt you when updates are available. You can check your current version with:
+
+```bash
+openhands --version
+```
+
+## CLI Commands and Settings
 
 The OpenHands CLI provides various commands for managing settings, conversations, and configurations. For detailed information about available commands, configuration options, and troubleshooting, see the [CLI Settings](/openhands/usage/run-openhands/cli-settings) page.
 
 Key features include:
 - Interactive settings configuration with `/settings`
 - MCP server management with `/mcp`
 - Conversation management with `/new` and `/help`
+- Memory management with `/condense`
 - Real-time status monitoring and control

openhands/usage/run-openhands/cli-settings.mdx

@@ -8,7 +8,25 @@ description: Configuration and command reference for the OpenHands Command-Line
 The OpenHands CLI provides commands to help with settings and conversations. These commands work by typing `/`
 followed by the command. To see a list of all the available commands run `/help` when running the CLI.
 
-### Settings and Configuration
+### Core Commands
+
+- `/help` - Display all available commands and their descriptions
+- `/new` - Start a new conversation without restarting the CLI
+- `/settings` - Configure your model, API key, and other preferences
+- `/mcp` - View MCP server status and configuration
+- `/condense` - Force compact conversation history to save memory and tokens
+
+### ACP Integration Commands
+
+<Note>
+ACP (Agent Computer Protocol) integration provides enhanced agent capabilities and is available starting from CLI version 1.3.0.
+</Note>
+
+When ACP is enabled, additional slash commands become available for managing agent sessions and interactions. These commands provide advanced control over agent behavior and session management.
+
+## Settings and Configuration
+
+### Interactive Settings Configuration
 
 You can update your model, API key, and other preferences using the `/settings` command. Just
 follow the prompts:
@@ -17,6 +35,23 @@ follow the prompts:
 - **Advanced settings**: Set custom models and toggle
   [memory condensation](/openhands/usage/settings/llm-settings#memory-condensation).
 
+### Memory Management
+
+The `/condense` command allows you to manually trigger conversation history condensation:
+
+- **Purpose**: Reduces token usage by compacting conversation history
+- **When to use**: When conversations become long or you want to optimize performance
+- **Effect**: Preserves important context while reducing overall message count
+
+### Confirmation and Approval Modes
+
+The CLI supports different modes for handling agent actions:
+
+- **Interactive Mode** (default): Prompts for approval before executing actions
+- **Always Approve Mode** (`--always-approve`): Automatically approves all actions
+- **LLM Approve Mode** (`--llm-approve`): Uses the LLM to intelligently approve actions
+- **Confirmation Mode**: Enhanced approval system with detailed action previews
+
 ### MCP Server Management
 
 <Note>

openhands/usage/run-openhands/headless-mode.mdx

@@ -6,6 +6,37 @@ description: You can run OpenHands with a single command, without starting the w
 
 This is different from [the CLI](/openhands/usage/run-openhands/cli-mode), which is interactive, and better for active development.
 
+## With CLI
+
+The OpenHands CLI supports headless mode through the `--headless` flag, which enables non-interactive execution with automatic exit functionality. This is perfect for scripting and automation scenarios.
+
+### Example Usage
+
+```bash
+# Launch a specific task into the terminal UI from headless mode
+openhands --headless --task "Fix the bug in main.py"
+
+# Run with experimental UI and a file in headless mode
+openhands --exp --headless --file src/app.py --task "Add error handling to this file"
+
+# Run in headless mode with auto-approval
+openhands --exp --headless --always-approve --task "Run tests and fix any failures"
+
+# Exit without confirmation prompts
+openhands --exp --headless --exit-without-confirmation --task "Generate documentation for the project"
+```
+
+### CLI Headless Options
+
+When using the CLI in headless mode, you can combine the `--headless` flag with other CLI options:
+
+- `--task "your task"`: Start with an initial task message
+- `--file path/to/file`: Start with a specific file context
+- `--always-approve`: Automatically approve all agent actions
+- `--llm-approve`: Use the LLM to automatically approve actions
+- `--exit-without-confirmation`: Exit without confirmation prompts
+- `--exp`: Enable experimental UI features
+
 ## With Python
 
 To run OpenHands in headless mode with Python: