Phase 1 — Discovery

Phase 1 is where the migration begins. The Discovery Orchestrator analyzes your source Sitecore XP site — codebase, content architecture, templates, and layout patterns — and produces the artifact foundation that every downstream phase depends on. Nothing else can start until Discovery is complete.

Duration: ~5–15 minutes | Orchestrator: Discovery Orchestrator

What Discovery does

Discovery runs two analysis streams in parallel:

Code analysis provisions a sandbox, checks out your source codebase, and produces:

A full component inventory with priority ordering (home page first)
Dependency graph across components
Style system analysis
Detection of any C# Rendering Contents Resolvers that will need adapter conversion in Phase 5

Content analysis connects to your source XP instance and produces:

Full template inventory
Item counts (used to calculate timeouts in Phase 3)
SXA readiness assessment
Layout pattern analysis
Serialization module configuration for Phase 3

Once both complete, the Discovery Orchestrator calls a third agent — Discovery Synthesis — which consolidates everything into a single, structured report and passes it to the workspace as artifacts.

Before you start

Make sure your workspace is fully configured:

Source XP connection verified (green on all four services)
Target SitecoreAI CMS (XMC) connection verified
Source and target Git connections active
Source site name confirmed — this is the Sitecore site name, not the repo name
A project created inside your workspace

⚠️
If any connector is missing or misconfigured, Discovery will surface it immediately. Use migration-get-info to check connector status and migration-set-connectors to fix it before proceeding.

How to run Discovery

Open your workspace in Kajoo Agentic
Create a new project inside the workspace (or open an existing one)
Open a new chat session inside the project
Type "Start Discovery" to invoke the Discovery Orchestrator
The agent will check your workspace configuration and confirm connectors
Review and approve the configuration — connectors, site name, repo paths
Optionally provide SitecoreAI CMS (XMC) environment details (Edge Context ID, Editing Secret) — or skip for now, these are only needed before Phase 5
The agent kicks off parallel code and content analysis
Review the discovery report when complete

📘
A workspace can have multiple projects, and each project can have multiple chat sessions. For this migration recipe, one project per migration is the recommended approach. You can open additional chat sessions within the same project to run Phase 2 and Phase 3 in parallel.

What you'll decide

Decision	When
Reuse existing sandbox or start fresh	At the start, if a prior sandbox exists
Approve workspace configuration	Before analysis begins — connectors, site name, repo paths
Provide SitecoreAI CMS (XMC) environment details	Optional — skip if not ready, required before Phase 5
Code resume path	If existing code is found in the target branch (see below)
Content / SXA bootstrap	If content or SXA structure is already present in the target (see below)

Resuming previous work

Discovery detects prior work automatically so you don't repeat completed steps.

Code resume

If existing project code is found in the target branch (package.json, src/components, etc.), the agent asks how to proceed:

Use this code (resume migration) — reconciles source components against what's already migrated, builds a migration-state with completed/pending lists. Phase 2 picks up from the first pending component.
Start fresh on new branch — creates a new branch and scaffolds from scratch.
Fix config — wrong repo/branch — go back and correct the configuration before proceeding.

Content bootstrap

If content is already serialized and present in SitecoreAI CMS (XMC), Discovery will ask you to confirm. Confirming satisfies Phase 4's dependency on content migration — Phase 3 can be skipped.

SXA bootstrap

If the target site already has headless SXA structure (tenant, adoption, conversion, page designs), Discovery will ask you to confirm. Confirming satisfies Phase 5's dependency on SXA conversion — Phase 4 can be skipped.

What Discovery produces

Three artifacts are written to your workspace at the end of Discovery. Every downstream phase reads from these.

Artifact	What it contains
`code-discovery-report`	Component count and priority list, dependency graph, style system analysis, resolver inventory
`content-discovery-report`	Template inventory, item counts, SXA readiness assessment, layout patterns
`serialization-config.json`	Module definitions used by Phase 3 to set up content serialization

Discovery also establishes the workspace migration config — migration type, connector IDs, source site name, repo paths. View it anytime with migration-get-info. Update it with migration-set-connectors.

Checking progress

Call migration-workflow-status at any time to see the pipeline state — which phases are complete, in progress, or blocked, and what needs to happen next.

Troubleshooting

Issue	Fix
"Source/target connector not configured"	Use `migration-set-connectors` to configure, `migration-get-info` to verify
"Source site name validation failed"	Site name ≠ repo name. Use `migration-set-connectors` with `source.siteName` to correct
Discovery artifacts required (in a later phase)	Run Discovery first — all downstream phases depend on its output

What's next

Once Discovery is complete, Phases 2 and 3 can start — and they can run in parallel since they operate on independent systems.


Start code migration	Continue to Phase 2 — Code Migration
Start content migration	Continue to Phase 3 — Content Migration
Run both in parallel	Open a second chat session within the same project for Phase 3 while Phase 2 runs — they operate on independent systems