Triple-C/container/mission-control/README.md

Flight Control

An AI-first software development lifecycle methodology using aviation metaphors to bridge human intent and AI execution.

What is Mission Control?

This repository is a centralized command center for managing multiple projects in parallel. Each project may have its own stack, systems, and constraints, but mission-control provides a consistent workflow and orchestration layer across all of them.

• Project registry — Track active projects with paths, remotes, and configurations
• Shared methodology — Apply structured planning regardless of project differences
• Claude Code skills — Interactive tools for mission, flight, and leg creation

Artifacts (missions, flights, legs) are created in target projects, not here. Mission-control holds the methodology, skills, and coordination—your projects hold the work.

The Aviation Model

Flight Control organizes work into three hierarchical levels, each optimized for its primary audience:

Mission (human-optimized)
  └── Flight (balanced)
        └── Leg (AI-optimized)

• Missions define outcomes in human terms—what success looks like and why it matters
• Flights translate outcomes into technical specifications with planning checklists
• Legs provide structured, specific instructions optimized for AI consumption

Why Aviation?

Aviation succeeds through layered planning and clear handoffs. Pilots follow flight plans but improvise when conditions demand it—weather, emergencies, ATC instructions. Structured planning enables effective improvisation by providing a baseline to deviate from and return to. Similarly, Flight Control separates strategic intent (missions) from tactical execution (legs), with flights serving as the translation layer.

Agentic Workflow

LLM orchestrators: Run /agentic-workflow to drive multi-agent flight execution with Claude Code. The skill orchestrates the full leg cycle — design, implement, review, commit — using three separate Claude instances.

Getting Started

Prerequisites

• Claude Code installed
• A project on disk with a git remote, initialized with Claude Code (claude /init)

Walkthrough

1. Clone mission-control — Clone this repo and open it in Claude Code.

2. Set up the projects registry — Run /init-mission-control (or manually copy projects.md.template → projects.md and fill in your project details). This creates the central registry that all skills read from.

3. Initialize your project — Run /init-project and select your project. This creates .flightops/ in your target project with artifact configuration, methodology reference, and crew definitions.

4. Review agent crew files — Check the files in {project}/.flightops/agent-crews/. These define the crew composition (roles, models, prompts) for each phase. Customize them to your needs.

5. Create a mission — Run /mission. This interviews you about desired outcomes and creates a mission artifact in your target project.

6. Design a flight — Run /flight to break the mission into a technical specification with pre/in/post-flight checklists.

7. Execute — Run /agentic-workflow to drive multi-agent implementation. This orchestrates design, implement, review, and commit cycles across legs.

8. Debrief — Run /flight-debrief and /mission-debrief after completion to capture lessons learned.

Documentation

1. Overview — Philosophy and principles behind Flight Control
2. Missions — Writing outcome-driven mission statements
3. Flights — Creating technical specifications with pre/post checklists
4. Flight Logs — Recording execution progress and decisions
5. Legs — Structuring AI-optimized implementation steps
6. Workflow — End-to-end flow from mission to completion

Core Concepts

The Audience Gradient

Documentation becomes progressively more structured as it moves down the hierarchy:

Level	Audience	Style
Mission	Humans, stakeholders	Narrative prose, outcome-focused
Flight	Developers, AI	Technical spec with checklists
Leg	AI agents	Structured format, explicit criteria

Lifecycle States

Each level tracks progress through defined states:

• Missions: planning → active → completed (or aborted)
• Flights: planning → ready → in-flight → landed → completed (or aborted)
• Legs: planning → ready → in-flight → landed → completed (or aborted)

Scaling

Flight Control scales from solo developers to teams. The methodology provides structure and continuity across sessions regardless of team size.

Artifact Organization

The hierarchy nests naturally:

Mission
├── Mission Debrief
└── Flight
    ├── Flight Log
    ├── Flight Briefing
    ├── Flight Debrief
    └── Leg

How you store these artifacts depends on your project's needs. Flight Control supports multiple artifact systems:

• Markdown files — Version-controlled documentation in your repository
• Issue trackers — Jira, Linear, GitHub Issues with linked relationships
• Hybrid — Missions in markdown, flights/legs as tickets

Each project configures its artifact system during initialization. The methodology and Claude Code skills adapt to your choice.

Claude Code Skills

Flight Control includes Claude Code skills for interactive planning:

Skill	Purpose
/init-mission-control	Set up the projects registry
/init-project	Initialize a project for Flight Control
/mission	Create outcome-driven missions through research and interview
/flight	Create technical flight specs from missions
/leg	Generate implementation guidance for LLM execution
/flight-debrief	Post-flight analysis for continuous improvement
/agentic-workflow	Drive multi-agent flight execution
/mission-debrief	Post-mission retrospective for outcomes assessment
/daily-briefing	Cross-project status report with health assessment

Recommended Workflow

All work runs from a single Mission Control session. Mission Control handles planning directly and spawns agents into the target project's context for implementation, review, and commits. Each spawned agent gets a clean context with only the information it needs, while Mission Control maintains continuity across the entire flight.

Context Strategy

• Mission Control: Long-running session spanning an entire flight — accumulates knowledge across legs, orchestrates all work
• Spawned agents: Fresh context per task — designed with precise instructions and the relevant artifacts, execute in the target project directory

Claude Code's version control in mission-control acts as the orchestrator for development of the remote project. No second interactive session is needed.

The Cycle

sequenceDiagram
    participant MC as Mission Control
    participant A as Spawned Agents

    Note over MC,A: ─── Mission Planning ───
    MC->>MC: /mission — research, interview, define outcomes
    MC->>MC: Review and confirm mission

    Note over MC,A: ─── Flight Planning ───
    MC->>MC: /flight — create technical spec, checklists
    MC->>MC: Review and confirm flight

    Note over MC,A: ─── Execution ───
    MC->>MC: /agentic-workflow

    loop For each leg
        Note over MC: Design phase
        MC->>A: Spawn designer agent
        A->>A: Design leg spec
        A-->>MC: Leg designed

        MC->>MC: Review leg design

        Note over MC: Implement phase
        MC->>A: Spawn implementer agent<br/>(target project context)
        A->>A: Implement leg, update logs
        A-->>MC: Implementation complete

        Note over MC: Review phase
        MC->>A: Spawn reviewer agent<br/>(target project context)
        A->>A: Review changes, verify criteria
        A-->>MC: Review complete

        Note over MC: Commit phase
        MC->>A: Spawn commit agent<br/>(target project context)
        A->>A: Stage and commit changes
        A-->>MC: Committed

        MC->>MC: Update flight checklist
    end

    Note over MC,A: Flight lands

    Note over MC,A: ─── Debrief ───
    MC->>MC: /flight-debrief
    MC->>MC: /mission-debrief

Why This Matters

A single orchestrating session eliminates context drift between planning and execution. Mission Control sees every leg's outcome and carries that knowledge forward into the next design. Spawned agents get clean, focused contexts — they don't need flight-wide memory because Mission Control provides exactly the context they need. Artifacts stay synchronized because one session owns the full lifecycle.

License

MIT