Guren DocsGuren Docs
Guide/guides

Glossary

Short definitions for common terms in the Guren docs.

Glossary

Short definitions for common terms in the Guren docs.

Stack

  • Bun: JavaScript/TypeScript runtime used to run and build the app.
  • Hono: Lightweight HTTP server; Guren builds its routing and controllers on top.
  • Inertia.js: Bridges server responses and SPA navigation by sending page props.
  • React: UI library used for pages.
  • Vite: Frontend dev server and build tool.
  • Drizzle ORM: Type-safe ORM; Guren models connect to Drizzle schemas.

App structure

  • MVC: Model, View, Controller separation.
  • Route: A path + HTTP method mapping to a controller or handler.
  • Controller: Class that handles requests and returns responses.
  • Model: Class tied to a database table via static table.
  • View: React page in resources/js/pages/.
  • Middleware: Code that runs before or after a request.
  • Provider: Boot-time registration for services and configuration.
  • Context: Hono Context with request and response helpers.

Database

  • ORM: Maps tables to classes.
  • Schema: Table definitions in db/schema.ts.
  • Migration: SQL files that apply schema changes.
  • Seeder: Script to insert sample data.
  • Database URL: Connection string in .env as DATABASE_URL.
  • RQB (Relational Query Builder): Drizzle query builder for joins and aggregates.
  • Eager load: Fetch related records in one go (e.g. with()).

Frontend

  • SPA: Single page app navigation without full reloads.
  • SSR: Server-side rendering for the first HTML response.
  • Props: Data passed to React components.
  • HMR: Hot reload of frontend changes during development, handled by Vite. Backend changes (controllers, routes, models) reload too, since the dev server runs under bun --hot.
  • Inertia page: A React component referenced by this.inertia(...).

CLI

  • create-guren-app: CLI to scaffold a new app.
  • guren CLI: bunx guren make:* commands for generators and tooling.

Agent harness

  • Agent harness: the files create-guren-app installs for coding agents: what an agent reads first, what runs after it edits, and what runs before it ends a turn.
  • CLAUDE.md: the project guide Claude Code reads at the start of a session. Claude Code docs
  • Rules: instructions in .claude/rules/, loaded only when the agent edits a file their paths match. Claude Code docs
  • Skills: procedures the agent follows for a kind of task, in .claude/skills/<name>/SKILL.md. Claude Code docs
  • Subagents: agents with their own instructions and context that the main agent invokes, in .claude/agents/. Claude Code docs
  • Hooks: commands Claude Code runs at points such as session start, a file edit, or the end of a turn, configured in .claude/settings.json. Claude Code docs
  • Gate (guren gate): runs codegen, typecheck, lint, check, audit and the tests in one command. CI and the Stop hook run the same one.

Plan states

plan:status, plan:verify and plan:next print these values for steps and elements. See Implementation Plans for the details.

  • verified: every verify command passed, and the record still holds.
  • failed: a verify command failed; there is something to fix.
  • blocked: the environment kept a command from running (a missing script, an unreachable database, a timeout). Not a failed implementation.
  • drifted: a file the record fingerprinted changed after it verified. Re-check it with plan:verify --step.
  • stalled: the Stop hook sent the agent back and the step still did not verify, so it recorded why and let the stop through. The next plan:next reports it.
  • held: something the step depends on changed in the app after approval, so plan:next holds the step.
  • waived: accepted incomplete with plan:waive.
  • advisory: a result that is shown but never fails --ci or guren gate.

Start here