multi-repo

# Multi-Repository Orchestration Skill

## Purpose

Coordinate atomic changes across multiple repositories when features span repo boundaries. Track dependencies, manage linked PRs, and detect breaking changes before they propagate.

## When to Use This Skill

**USE FOR:**

- Changes that require coordinated updates across multiple repositories
- Managing repository dependency graphs
- Creating linked PRs that should merge together or in sequence
- Detecting breaking changes that affect dependent repositories
- Microservices or monorepo-alternative architectures

**AVOID FOR:**

- Single-repository changes (use default workflow)
- Forking/syncing operations (use git directly)
- Documentation-only changes
- Changes with no cross-repo dependencies

## Storage Location

All multi-repo data is stored in `.claude/data/multi-repo/`:

- `dependencies.yaml` - Repository dependency graph
- `linked-prs.yaml` - Currently active linked PR sets
- `breaking-changes.log` - History of detected breaking changes

## Dependency Graph Format

The dependency graph uses simple YAML:

```yaml
# .claude/data/multi-repo/dependencies.yaml
version: "1.0"
repositories:
  my-org/api-server:
    depends_on: []
    exposes:
      - type: api
        name: REST API v2
        contract: openapi.yaml

  my-org/web-client:
    depends_on:
      - repo: my-org/api-server
        type: api
        version: ">=2.0"
    exposes: []

  my-org/mobile-app:
    depends_on:
      - repo: my-org/api-server
        type: api
        version: ">=2.0"
    exposes: []
```

## Core Operations

### Operation 1: Initialize Dependency Graph

**When:** Setting up multi-repo tracking for the first time

**Process:**

1. Create `.claude/data/multi-repo/` directory if not exists
2. Create initial `dependencies.yaml` with current repository
3. Prompt for known dependencies (repos this one depends on)
4. Prompt for known dependents (repos that depend on this one)
5. Save and display the graph

**Command:** "Initialize multi