reins

Stop the Reins orchestrator UI.

Launch the Reins sprint orchestrator. Opens a mission-control UI for multi-agent sprint management with live streaming, interrupt, and direct agent chat.

Independent code reviewer using GPT-5.4. Triggered immediately after PR creation, runs in parallel with bot comments. Provides a second-opinion review isolated from the primary Claude reviewer.

UX/UI designer. Generates HTML mockup variants for frontend tasks.

Implementation specialist. Writes code and tests in a git worktree using Codex GPT 5.2. Follows TDD, SOLID, Clean Code. Creates PR when done.

Specification writer. Writes SPEC only (not plans). Explores codebase, runs reviews, produces structured specifications with AC categories and Quality Ratchet.

Project manager. Arranges iteration from Planner spec: computes card dependencies, assigns wave numbers, writes planning-with-files output (task_plan.md). Also serves as PR quality gate — reads all raw comments and decides pass/fail.

Invoke before writing any backend Python code. Enforces Red-Green-Refactor cycle with project-specific Clean Code constraints.

Invoke before writing any frontend React/TypeScript code. Enforces component-driven TDD with project-specific constraints.

Sprint planning only. Use when the user wants to plan a sprint without launching the full UI. Runs the Planner agent to generate a spec from GitHub issues.

Run API tests in an isolated microsandbox microVM with embedded Postgres. Use when tester agents need to start the full app and test API endpoints in an isolated environment. Each sandbox has its own network stack, DB, and app instance. Trigger on: "sandbox test", "test in sandbox", "run API tests isolated", "sandbox-tester".

Run a multi-agent sprint. Use when the user wants to plan, execute, review, and test a set of GitHub issues as a coordinated sprint.