Kiro CLI MCP Server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| KIRO_MCP_CLI_PATH | No | Path to kiro-cli executable | kiro-cli |
| KIRO_MCP_TASK_TTL | No | Task result TTL in seconds | 3600 |
| KIRO_MCP_LOG_LEVEL | No | Logging level | INFO |
| KIRO_MCP_POOL_SIZE | No | Process pool size | 5 |
| KIRO_MCP_LOG_RESPONSE | No | Log full CLI responses for debugging | true |
| KIRO_MCP_MAX_SESSIONS | No | Maximum concurrent sessions | 10 |
| KIRO_MCP_POOL_ENABLED | No | Enable process pooling | true |
| KIRO_MCP_DEFAULT_AGENT | No | Default agent to use | kiro_default |
| KIRO_MCP_DEFAULT_MODEL | No | Default AI model for kiro-cli | claude-opus-4.5 |
| KIRO_MCP_POOL_MAX_USES | No | Max uses per process before recycling | 100 |
| KIRO_MCP_POOL_IDLE_TIME | No | Process idle time before recycling in seconds | 300 |
| KIRO_MCP_COMMAND_TIMEOUT | No | Command timeout in seconds - IDE-optimized | 30 |
| KIRO_MCP_MAX_ASYNC_TASKS | No | Maximum concurrent async tasks | 100 |
| KIRO_MCP_SESSION_TIMEOUT | No | Session idle timeout in seconds | 300 |
| KIRO_MCP_CLEANUP_INTERVAL | No | Session cleanup check interval in seconds | 30 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Server capabilities have not been inspected yet.
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| kiro_chatC | Send a chat message to kiro-cli and get AI response |
| kiro_session_createB | Create a new kiro-cli session. If working_directory is not provided or does not exist, the current directory will be used. |
| kiro_session_listB | List all active kiro-cli sessions |
| kiro_session_switchC | Switch to a specific session |
| kiro_session_endC | End a kiro-cli session |
| kiro_commandC | Execute a kiro-cli command (e.g., /mcp, /help) |
| kiro_agents_listB | List available custom agents |
| kiro_historyC | Get conversation history for a session |
| kiro_history_clearC | Clear conversation history for a session |
| kiro_chat_asyncA | Start an async chat task for streaming-like experience. Use kiro_task_status to poll for results. |
| kiro_task_statusA | Get status and partial results of an async task. Use for polling streaming results. |
| kiro_task_cancelC | Cancel a running async task |
| kiro_task_listC | List active async tasks |
| kiro_pool_statsB | Get process pool statistics for performance monitoring |
| kiro_session_clearA | Clear kiro-cli session history in working directory (deletes .kiro/session.json) |
| kiro_session_saveC | Save current kiro-cli session to a file using /save command |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| Active Sessions | List of all active kiro-cli sessions |
| Available Agents | List of configured custom agents |
| Server Configuration | Current server configuration (read-only) |
TDQS
Scored across 16 tools
Most tools have distinct purposes, but there is some overlap between kiro_chat and kiro_command, as both involve sending messages or commands to kiro-cli, which could cause confusion. However, descriptions clarify that kiro_chat is for AI responses and kiro_command for CLI commands, mitigating ambiguity.
All tool names follow a consistent kiro_ prefix with snake_case and clear verb_noun patterns (e.g., kiro_agents_list, kiro_session_create). This uniformity makes the tool set predictable and easy to navigate.
With 16 tools, the count is slightly high but reasonable for a CLI server covering agents, chat, sessions, and tasks. It feels comprehensive without being overwhelming, though some tools like kiro_session_clear and kiro_history_clear might be redundant.
The tool set provides good coverage for managing kiro-cli sessions, agents, chat, and async tasks, with clear CRUD-like operations (e.g., create, list, end, clear). Minor gaps exist, such as no direct tool for updating sessions or agents, but agents can work around this using chat or commands.