claude-relais

Orchestrate Claude Code tasks and Cursor builds with plan-build-verify contracts.

4|Updated Feb 11, 2026
One-click install
npx skills add https://github.com/clementrog/claude-relais --skill claude-relais
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: claude-relais
Source: https://github.com/clementrog/claude-relais/tree/main
Command: npx skills add https://github.com/clementrog/claude-relais --skill claude-relais

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires ajv, commander, micromatch, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Claude-based orchestration often requires coordinating planning, execution, and verification across tools; claude-relais provides a router-first workflow that keeps heavy reasoning in Claude and delegates deterministic builds to Cursor with strict safety contracts.

Core Features & Use Cases

  • Router-first orchestration: fresh repos trigger guided onboarding, existing repos get next-step options, and BUILD is always dispatched to Cursor.
  • Deterministic BUILD: Cursor headless execution ensures throughput and traceable changes.
  • Safety contracts: GO/PLAN -> BUILD -> VERIFY with git-truth validation and scope guardrails.

Quick Start

Ask Claude to start the orchestration for your repo and let claude-relais onboard or continue while Cursor builds changes.

Frequently Asked Questions about claude-relais

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate Claude Code orchestration with deterministic builds and verification?

Claude Code orchestration with deterministic builds is automated using a router-first workflow that keeps reasoning in Claude, dispatches headless Cursor builds, and verifies changes against git truth with strict safety contracts.

What is router-first orchestration for AI-driven code generation?

Router-first orchestration for AI-driven code generation routes each task tick through guided onboarding for fresh repos or continuation logic for existing repos, dispatching deterministic builds to Cursor while Claude handles heavy reasoning.

How do I enforce plan-build-verify contracts for automated code changes?

Plan-build-verify contracts for automated code changes are enforced by routing GO/PLAN stages to Claude, dispatching BUILD to Cursor headless execution, and running VERIFY against git-truth validation with scope guardrails.

Can I use Cursor headless execution with Claude Code for ongoing repository automation?

Yes, Cursor headless execution integrates with Claude Code for ongoing repository automation by routing existing repos through next-step continuation options while delegating deterministic builds to Cursor with traceable changes.

Do I need both Claude and Cursor to run automated orchestration with guardrails?

Yes, automated orchestration with guardrails requires Claude for heavy reasoning and planning, and Cursor for deterministic headless builds, ensuring throughput and traceable changes validated against git truth.

What are the limitations of router-first orchestration for team automation workflows?

Router-first orchestration limitations include strict dependency on the plan-build-verify contract flow, requiring both Claude and Cursor environments, and adherence to scope guardrails that may restrict unconstrained automated changes.