Reference on vshuraeff — Infrastructure. DevOps. Cloud.

GARDEN principles

Mon, 01 Jan 0001 00:00:00 +0000

About this document

GARDEN is a set of six language-agnostic software design principles for the agent-first era, in which code is written, reviewed, and analyzed primarily by LLM agents while humans set intent and constraints but rarely read the code line by line. This document is the canonical, self-contained statement of the six principles. It does not depend on any other file in this documentation set and may be copied into any project as a standalone agent-facing reference.

GARDEN compliance checklist

Mon, 01 Jan 0001 00:00:00 +0000

How to use this checklist

This checklist restates the rules from the GARDEN principles reference as discrete, checkable items. Each item is tagged [MUST] or [SHOULD] and carries the rule it maps to.

For an agent reviewer: evaluate each item against the code under review, and cite file-and-line evidence for every answer, whether the item passes or fails. Do not mark an item as passing without evidence; an unverifiable item is a finding, not a pass. Report findings as hypotheses per How to review code as an agent, not verdicts.

GARDEN glossary

Mon, 01 Jan 0001 00:00:00 +0000

Terms below are used with exactly the meaning given here throughout this documentation set. Each entry links to the principle section in GARDEN principles where the term is normative.

agent — An LLM-driven program that reads, writes, and reviews code via tools. Normative across all six principles; see GARDEN principles.

canonical name — The single name a concept has across the codebase; synonyms break grep-first discovery. Normative in G — Grep-first Discoverability.

context file — A short hand-maintained instruction file for agents (for example, AGENTS.md or CLAUDE.md) loaded at session start. Normative in N — Navigable Knowledge.