Documentation

Build with Confidence

Everything you need to create, deploy, and scale AI agents. From quickstart guides to full API reference.

Quick Start

Define agents with simple markdown files. Configure personality, tools, guardrails, and behavior — all in one place.

agents/calliope.md
---
name: Calliope
role: Content Writer
model: gpt-4o
temperature: 0.7
max_tokens: 4096
tools:
  - web_search
  - grammar_check
  - plagiarism_scan
guardrails:
  - no_hallucination
  - brand_voice_check
---

# Calliope — Content Writer

## Personality
You are Calliope, a seasoned content writer with 10 years
of experience in SaaS marketing. You write in a clear,
conversational tone that balances authority with approachability.

## Instructions
- Always research the topic before writing
- Include data points and statistics where possible
- Structure content with clear headings and short paragraphs
- End every piece with a compelling CTA
- Match the brand voice guide in `/docs/brand-voice.md`

## Examples
**Good:** "AI agents reduced manual QA time by 73% — here's how."
**Bad:** "AI is really cool and you should use it."
crew.yaml
# crew.yaml — Orchestrate your agent squad

name: content-pipeline
description: End-to-end content production workflow

agents:
  - agent: agents/odin.md        # Researcher
    step: 1
    task: "Research the topic and produce a brief"

  - agent: agents/calliope.md    # Writer
    step: 2
    depends_on: [1]
    task: "Write a blog post based on the research brief"

  - agent: agents/apollo.md      # SEO Specialist
    step: 3
    depends_on: [2]
    task: "Optimize the draft for target keywords"

  - agent: agents/athena.md      # Squad Lead (reviewer)
    step: 4
    depends_on: [3]
    task: "Final review — quality, tone, accuracy"

settings:
  parallel: false          # Run steps sequentially
  max_retries: 2           # Retry failed steps
  timeout: 300s            # Per-step timeout
  output_dir: ./output     # Where to save results
  notify:
    on_complete: slack://content-team
    on_failure: email://admin@company.com
terminal
# Install the CLI
$ npm install -g @missioncontrolhq/cli

# Initialize a new project
$ mchq init my-agent-squad
  āœ” Created project structure
  āœ” Generated sample agents/
  āœ” Created crew.yaml

# Project structure:
my-agent-squad/
ā”œā”€ā”€ agents/
│   ā”œā”€ā”€ athena.md        # Squad Lead
│   ā”œā”€ā”€ calliope.md      # Writer
│   ā”œā”€ā”€ odin.md          # Researcher
│   └── apollo.md        # SEO
ā”œā”€ā”€ crew.yaml            # Orchestration config
ā”œā”€ā”€ tools/
│   └── custom-search.js # Custom tool definitions
└── mchq.config.json     # API keys & settings

# Run a single agent
$ mchq run agents/calliope.md --input "Write a post about AI agents"
  ā ‹ Calliope is researching...
  ā ‹ Calliope is writing...
  āœ” Done — output saved to ./output/post.md

# Run the full crew pipeline
$ mchq crew run crew.yaml --input "AI agents for customer support"
  āœ” Step 1: Odin completed research brief
  āœ” Step 2: Calliope wrote the draft (1,847 words)
  āœ” Step 3: Apollo optimized — SEO score 94
  āœ” Step 4: Athena approved — no revisions needed
  āœ” Pipeline complete — output saved to ./output/

Need help?

Join our Discord community or reach out to support — we're here to help you succeed.

Join CommunityContact Support