Guide

Genome

The Genome is REAP's authoritative knowledge source — architecture principles, development conventions, technical constraints, and domain rules. It is the DNA of your project.

Structure

text

.reap/genome/
├── application.md     # Project identity, architecture, conventions
├── evolution.md       # AI behavior guide, evolution direction
└── invariants.md      # Absolute constraints (human-only edits)

Three Files

application.md — Project identity, architecture decisions, development conventions, and constraints.
evolution.md — AI behavior guide, interaction principles, code quality rules, and soft lifecycle rules.
invariants.md — Absolute constraints that can never be violated. Only the human can modify this file.

Genome Immutability

The Genome is never modified during a normal generation. Issues discovered during Implementation are recorded as genome-change backlog items and applied only during the Completion adapt phase.

Exception: Embryo generations (early-stage projects) allow free genome modification during any stage. Once the project matures, the AI proposes transitioning from embryo to normal during the adapt phase, and the human approves.

Always Loaded

All three genome files are fully loaded into the AI agent's context at session start via CLAUDE.md. The agent always has access to your project's architecture, conventions, and constraints — no manual briefing needed.

Evolution through Generations

At the end of each generation (Completion adapt phase), genome-change backlog items are reviewed and applied to the Genome. This ensures the Genome evolves deliberately, informed by what actually happened during the generation.

Knowledge Management

Use /reap.knowledge to review and manage the Genome and Environment. The command provides options to reload context, inspect genome files, and manage environment data.