Understanding Spec Kitty Missions

Divio type: Tutorial

Spec Kitty supports four mission types that tailor the workflow and artifacts to your goal.

Time: ~45 minutes Prerequisites: Completed Getting Started

Understanding Spec Kitty missions - Mission Kitty briefing

Stylized splash — the spy-themed cards are decorative. The four actual mission types (Software Dev, Deep Research, Documentation, Plan) are described below.

What Is a Mission?

A mission selects the default templates, prompts, and outputs for your work. You choose it during spec-kitty init or when running /spec-kitty.specify. A mission is the unit that runs the whole spec-to-merge lifecycle; each mission decomposes into work packages (WPs) your agent implements one at a time — see Your First Mission for that loop.

The Four Mission Types

For the full purpose, phases, and required artifacts of each type, see Mission Types — this page only covers what you need to pick one and get moving.

Software Dev Kitty

Best for building software features and products.

Example use cases:

  • New API endpoint
  • UI feature development
  • Performance improvements

Deep Research Kitty

Best for structured research deliverables.

Example use cases:

  • Competitive analysis
  • Architecture decision research
  • Technology evaluation

Documentation Kitty

Best for creating or updating documentation sets.

Example use cases:

  • End-user docs refresh
  • API reference overhaul
  • Internal playbooks

Plan Kitty

Best for goal-oriented planning and strategy documents where you expect to iterate — not code, not docs.

Example use cases:

  • Multi-quarter roadmap or strategy doc
  • A structured decision/investigation writeup with a rollback-to-draft cycle
  • Any goal-oriented plan you want gated by explicit review/approval rather than a code-review or publication gate

Try It: Create a Research Mission

Create a project, then start the research mission:

spec-kitty init my-research-project --ai claude
cd my-research-project

Use /spec-kitty.specify with the research mission to start a research workflow.

In your agent:

/spec-kitty.research Compare three task queue options for a Python service.

Expected results (abridged):

  • kitty-specs/###-task-queue-research/ directory
  • Research artifacts defined by the mission templates

How Missions Affect Your Workflow

  • Templates: Each mission uses its own spec/plan/templates.
  • Artifacts: Research missions create research notes; documentation missions generate Divio-oriented sections.
  • Validation: Review criteria differ based on mission expectations.

The Mission Lifecycle, Briefly

specify → plan → tasks → analyze → implement (per WP: claim → implement → review, with rejection cycles) → accept → merge

Every mission type — Software Dev, Deep Research, Documentation, and Plan — moves through this same loop; only the artifacts and templates change per type. For the full walkthrough with real commands, see Your First Mission.

Troubleshooting

  • "Unknown mission": Use spec-kitty mission list to list available missions.
  • Missing /spec-kitty.research: Your agent's command files are stale or weren't generated for this project. Run spec-kitty upgrade to regenerate them, then check ls .claude/commands/ (or the equivalent directory for your agent) for spec-kitty.research.md.

What's Next?

Explore the full workflow in Your First Mission or dive deeper into specific missions.

Reference Documentation

Edit this page