Phase 0: Prerequisites¶

Before you start planning, make sure the scaffolding is in place.

Checklist¶

• Your repo has an AGENTS.md explaining tool rules, model selection, and safety constraints
• Beads is initialized (or will be initialized) for the project
• You've decided where your canonical plan lives (e.g., PLAN.md in repo root)
• Agent coordination tooling is available (Agent Mail, if using multi-agent): Python | Rust (newer port, WIP)
• You have access to a strong reasoning model for plan critique (web app with extended thinking)

About AGENTS.md¶

AGENTS.md is the instruction manual agents read before doing anything. It should include:

• Tool rules -- which tools agents can use and how
• Model selection guidance -- which model to use for which task type
• Safety constraints -- what agents must never do (force push, delete prod data, etc.)
• Repository conventions -- commit message format, branch naming, test requirements
• Coordination rules -- how agents claim work and communicate (if multi-agent)

About Beads¶

Beads is the task management system. Think of it as a dependency-aware task graph where:

• Epics contain related groups of work
• Tasks are individual units of work
• Subtasks break tasks into smaller steps
• Dependencies define what must complete before something else can start

Each bead is self-documenting: it includes enough context that an agent can pick it up and work without re-reading the entire plan.

Setting Up the Flywheel¶

The easiest way to get all tooling configured is the Agentic Coding Flywheel Setup, which transforms a fresh VPS into a fully-armed agentic coding environment. The setup wizard is available at agent-flywheel.com.