Skip to content

Indigo Skills

Add a single reference file — SKILLS.md — to your AI agent’s context and it learns every Indigo CLI command, output format, error code, and multi-step workflow. Your agent can then authenticate, search signals, browse meetings, and run setup — all through natural language.

How it works

Your agent reads SKILLS.md once to understand the CLI, then runs indigo commands in a shell on your behalf. Commands are authenticated with your credentials and return structured JSON that your agent parses automatically.

Skills vs MCP

Indigo offers two ways to connect AI agents to your data. You can use either or both.

SkillsMCP
How it worksAgent runs CLI commands in a shellAgent calls query tools directly
Best forAutomation, setup, account ops, scriptingData queries, analytics, exploration
RequiresAny agent with shell accessMCP-compatible client (Claude Desktop/Code)
Data accessCLI output (JSON or human-readable)Direct database queries with filters
SetupAdd SKILLS.md to agent contextRun indigo setup mcp

Setup

  1. Get the SKILLS.md file

    SKILLS.md is a single markdown file (~1,100 lines) that documents every Indigo CLI command — flags, output schemas, exit codes, and workflows.

    If you have access to the indigo-nx repository, copy SKILLS.md from the project root:

    Terminal window
    cp /path/to/indigo-nx/SKILLS.md ./SKILLS.md
  2. Add SKILLS.md to your agent

    Add SKILLS.md as a Project Knowledge file:

    1. Open Claude Desktop and create or open a project
    2. Click the project settings icon
    3. Under Project Knowledge, click Add file
    4. Select your SKILLS.md file

    Claude will include the file in every conversation within that project.

  3. Verify it works

    Ask your agent:

    Check my Indigo auth status.

    Your agent should run indigo auth status --json and report whether you’re authenticated. If it doesn’t recognize the command, check that SKILLS.md is loaded — some agents need a restart to pick up new files.

What your agent can do

Once SKILLS.md is loaded, your agent understands every command group in the Indigo CLI.

Authentication

Your agent can check auth status, guide you through login, and detect expired tokens.

Try asking:

Am I logged into Indigo?
Help me log in to Indigo.

For command details, see indigo auth.


Signals & insights

Your agent can search, filter, and retrieve meeting signals — decisions, actions, accomplishments, and key facts.

Try asking:

Find all action items from last week.
What decisions were made about the product launch?
Show me key facts from yesterday's meetings.
Signal typeDescription
decisionA decision made during a meeting
actionAn action item assigned to a person or team
accomplishmentA completed milestone or achievement
key_factAn important piece of information shared

For command details, see indigo signals.


Meetings

Your agent can browse your calendar, search by title or participant, and view meeting details with summaries.

Try asking:

What meetings do I have today?
Find meetings with Sarah from last week.
Show me the summary of the product planning meeting.

For command details, see indigo meetings.


Setup & configuration

Your agent can run the setup wizard, check configuration status, connect your calendar, manage API keys, and configure MCP.

Try asking:

Check my Indigo setup status.
Help me connect my Google Calendar.
Set up MCP for Claude Desktop.

For command details, see indigo config.


Multi-step workflows

Your agent chains commands together for complex tasks — the real power of Skills.

Daily briefing:

Prepare me for today's meetings.

Action item review:

What action items are pending from this week?

Meeting intelligence:

Search across my meetings and signals for anything about Q2 planning.

First-time setup:

Help me set up Indigo from scratch.

Weekly recap:

Give me a summary of what happened in my meetings this week.

What Skills don’t cover

Next steps

CLI commands

Full reference for every Indigo CLI command. CLI overview

MCP server

Give your agent direct database access as a complement to Skills. MCP overview

CLI automation

Scripting patterns, exit codes, and CI/CD integration. Automation

MCP use cases

Practical prompts for data queries via MCP. Use cases