MCP Setup

What is MCP?

The Model Context Protocol (MCP) is an open standard that allows AI assistants to interact with external tools and data sources. By running rmbr as an MCP server, AI assistants like Claude can directly create, query, and manage your work entries -- todos, goals, kudos, TILs, study topics, Slack messages, and tags -- all through natural conversation.

Starting the MCP Server

rmbr includes a built-in MCP server that runs over the stdio transport:

rmbr mcp

This starts the server and listens for MCP messages on stdin/stdout. You do not run this manually -- instead, configure your AI client to launch it automatically.

Claude Desktop Configuration

Add the following to your claude_desktop_config.json file:

{
  "mcpServers": {
    "rmbr": {
      "command": "rmbr",
      "args": ["mcp"]
    }
  }
}

This works because bun install automatically registers the rmbr command on your PATH via bun link.

The config file location depends on your platform:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json
• Linux: ~/.config/Claude/claude_desktop_config.json

Claude Code Configuration

For Claude Code, add the MCP server to your .claude/settings.json:

{
  "mcpServers": {
    "rmbr": {
      "command": "rmbr",
      "args": ["mcp"],
      "type": "stdio"
    }
  }
}

This can be placed in your project-level .claude/settings.json or your global settings file.

Available MCP Tools

All rmbr modules expose their functionality as MCP tools. Each tool name is prefixed with rmbr_.

Todo

Tool	Description
rmbr_todo_create	Create a new todo from raw input. Accepts enrichment fields for single-step creation.
rmbr_todo_list	List todos with optional status, pagination filters.
rmbr_todo_get	Get a single todo by ID.
rmbr_todo_transition	Transition a todo to a new status (start, pause, done, cancel).
rmbr_todo_enrich	Enrich a todo with title, priority, due date, and goal link.
rmbr_todo_delete	Soft-delete a todo.
rmbr_todo_restore	Restore a soft-deleted todo.

Goals

Tool	Description
rmbr_goal_create	Create a new goal. Accepts enrichment fields for single-step creation.
rmbr_goal_list	List goals with optional status, quarter, year filters.
rmbr_goal_get	Get a single goal by ID.
rmbr_goal_transition	Transition a goal to a new status (activate, complete, abandon).
rmbr_goal_enrich	Enrich a goal with title, quarter, year, and KPIs.
rmbr_goal_add_star_narrative	Add a STAR narrative (Situation, Task, Action, Result) to a goal.
rmbr_goal_get_star_narratives	Retrieve all STAR narratives for a goal.
rmbr_goal_quarterly_review_data	Get aggregated data for a quarterly review.
rmbr_goal_save_quarterly_review	Save a generated quarterly review.
rmbr_goal_related	Get all entities linked to a goal (todos, kudos, study, slack).
rmbr_goal_delete	Soft-delete a goal.
rmbr_goal_restore	Restore a soft-deleted goal.

Kudos

Tool	Description
rmbr_kudos_create	Create a kudos entry. Accepts enrichment fields for single-step creation.
rmbr_kudos_list	List kudos with optional direction filter.
rmbr_kudos_get	Get a single kudos entry by ID.
rmbr_kudos_enrich	Enrich kudos with person, direction, summary, context, and goal link.
rmbr_kudos_delete	Soft-delete a kudos entry.
rmbr_kudos_restore	Restore a soft-deleted kudos entry.

TIL

Tool	Description
rmbr_til_create	Create a TIL entry. Accepts enrichment fields for single-step creation.
rmbr_til_list	List TILs with optional domain filter.
rmbr_til_get	Get a single TIL by ID.
rmbr_til_search	Full-text search across all TIL entries.
rmbr_til_domains	List all known domains.
rmbr_til_enrich	Enrich a TIL with title, content, domain, and tags.
rmbr_til_delete	Soft-delete a TIL entry.
rmbr_til_restore	Restore a soft-deleted TIL entry.

Study

Tool	Description
rmbr_study_create	Create a study topic. Accepts enrichment fields for single-step creation.
rmbr_study_list	List study topics with optional status and domain filters.
rmbr_study_get	Get a single study topic by ID.
rmbr_study_transition	Transition a study topic (start, done, park).
rmbr_study_add_note	Add a note to a study topic.
rmbr_study_add_resource	Add a resource URL to a study topic.
rmbr_study_next	Get the next queued study topic.
rmbr_study_enrich	Enrich a study topic with title, domain, and goal link.
rmbr_study_delete	Soft-delete a study topic.
rmbr_study_restore	Restore a soft-deleted study topic.

Slack

Tool	Description
rmbr_slack_ingest	Ingest a Slack message with optional channel, sender, and timestamp.
rmbr_slack_list	List messages with optional channel, processed, and sentiment filters.
rmbr_slack_get	Get a single Slack message by ID.
rmbr_slack_set_sentiment	Set sentiment on a message (positive, negative, neutral).
rmbr_slack_link_todo	Link a Slack message to a todo.
rmbr_slack_link_goal	Link a Slack message to a goal.
rmbr_slack_mark_processed	Mark a Slack message as processed.
rmbr_slack_delete	Soft-delete a Slack message.
rmbr_slack_restore	Restore a soft-deleted Slack message.

Tags

Tool	Description
rmbr_tag_entity	Add a tag to any entity (todo, kudos, goal, til, study, slack).
rmbr_untag_entity	Remove a tag from an entity.
rmbr_tag_list	List all tags.
rmbr_tag_get_entities	Get all entities with a given tag, optionally filtered by type.
rmbr_entity_tags	Get all tags for a specific entity.

Search

Tool	Description
rmbr_search	Search across all modules (todos, goals, kudos, TILs, study, slack).

Skills for Claude Code

rmbr bundles AI workflow skills that guide Claude through multi-step workflows. Install them with:

rmbr skill install

This creates /rmbr-* slash commands in Claude Code (e.g., /rmbr-weekly-standup). See the CLI Usage guide for the full list of available skills.

Single-Step Creation

All create tools accept enrichment fields directly alongside the raw input. This means an AI assistant can create a fully enriched entity in a single tool call rather than requiring separate create and enrich steps.

For example, rmbr_todo_create accepts not only the raw input text but also title, priority, due_date, and goal_id -- all in one call. This reduces round trips and makes AI interactions faster and more natural.

Example Interaction

Here is how a conversation with an AI assistant using rmbr might look:

You: I just fixed the login bug that was causing session timeouts.

The AI assistant recognizes this as a completed task and uses rmbr tools behind the scenes:

1. Creates a todo via rmbr_todo_create with the raw input and enrichment fields:

   {
     "raw_input": "Fixed the login bug causing session timeouts",
     "title": "Fix login session timeout bug",
     "priority": "high"
   }

2. Transitions it to done via rmbr_todo_transition:

   {
     "id": 1,
     "status": "done"
   }

3. Creates a TIL via rmbr_til_create to capture the learning:

   {
     "raw_input": "Session timeouts were caused by login flow not refreshing tokens",
     "title": "Login session timeout root cause",
     "domain": "authentication"
   }

AI: Done. I have recorded that as a completed todo and captured a TIL about the session timeout root cause. Would you like to link it to a goal?

This seamless interaction is possible because the AI assistant has direct access to all rmbr tools through MCP.