Fletcher Mission Control — AI skill for Claude Code

Fletcher Mission Control

**Lightweight dashboard for monitoring autonomous AI agents. Cost tracking, task board, scheduled jobs, and memory viewer. Built for OpenClaw on Mac mini M4.**

Created by Alex · March 2026

---

What This Is

A four-screen Next.js dashboard for monitoring and managing an autonomous AI agent running on [OpenClaw](https://openclaw.ai). Designed for local deployment inside the agent's VM, accessible via SSH port forwarding.

No external dependencies beyond Next.js and React. Data is stored in local JSON files — no database server required.

Screenshots

Cost Tracker

Track daily and monthly API spend across model providers. Budget alerts when spending exceeds thresholds. SVG spend-over-time chart.

Task Board

Kanban-style task management with four columns: Backlog → In Progress → Review → Done. Create tasks, move them between columns, and track what your agent is working on.

Scheduled Jobs

View and manage recurring tasks like heartbeat checks, morning briefs, and end-of-day logs. Pause and resume individual jobs.

Memory Viewer

Browse all workspace and memory markdown files with full-text search. Expandable file viewer with content preview.

Tech Stack

• Next.js (App Router) — React framework
• Vanilla CSS — Custom dark-mode design system, no Tailwind dependency
• JSON file storage — Zero-config data persistence in data/ directory
• Node.js fs — Direct filesystem access for reading agent memory files

Quick Start

1. Clone and install

git clone git@github.com:alexschar/fletcher-mission-control.git
cd fletcher-mission-control
npm install

2. Create seed data (optional)

mkdir -p data

cat > data/costs.json << 'EOF'
[
  {"id":"1","provider":"OpenAI","model":"gpt-5-nano","tokens_in":4500,"tokens_out":1200,"cost_est":0.0285,"notes":"heartbeat check","timestamp":"2026-03-14T08:00:00Z"},
  {"id":"2","provider":"Anthropic","model":"claude-sonnet","tokens_in":12000,"tokens_