This reference defines how to adapt domain pages for mixed-stack repositories.
Use this page when a business domain spans more than backend services alone.
Typical mixed-stack domain surfaces include:
- app screens
- H5 pages
- backend APIs
- Python tasks or workers
- bridge entrypoints
- callback handlers
The classic backend-oriented structure:
- domain
- service
- spec
is often too narrow for mixed-stack repositories.
In mixed-stack analysis, a reusable domain page should connect:
- business intent
- user-facing or system-facing entry surfaces
- technical systems and modules
- rules, contracts, and specs
Recommended structure:
# Domain Page
## 1. Domain Definition
- domain name
- current scope
- current boundary
## 2. Entry Surfaces
| surface type | name | path/module | role | evidence level | code locations |
| --- | --- | --- | --- | --- | --- |
## 3. Systems / Services / Modules
| unit | stack type | responsibility | upstream surfaces | downstream links | code locations |
| --- | --- | --- | --- | --- | --- |
## 4. Rules / Contracts / Specs
| rule or contract | applies to | status | evidence | code/doc locations |
| --- | --- | --- | --- | --- |
## 5. Current Evidence Boundary
- confirmed
- clue-level only
- unresolvedUseful surface labels include:
app-screenh5-pagebackend-apipython-taskbridge-entrycallback-handleradmin-pagecron-job
Only use labels supported by visible code evidence.
A mixed-stack domain page should not repeat low-level codemap pages.
Instead, it should elevate existing evidence into a reusable context layer.
This means:
- summarize the domain-facing surfaces
- connect them to systems or modules already identified
- connect those units to rules or contracts already visible
- keep unresolved parts explicit
Do not treat a domain page as a place to guess product intent.
Only write:
- what the code or repository structure directly supports
- what entry surfaces are visible
- what systems or modules are visibly linked
- what rules or specs are actually present
If domain intent is only partially visible, say so.
For mixed-stack repositories, domain pages should usually be built after:
- workspace layering
- subsystem codemap pages
- cross-layer router-map pages
- interface-mapping pages when needed
This keeps the domain page evidence-backed.
Use phrases like:
Domain context lifted from lower-level codemap evidenceEntry surface confirmed in codeSystem-module linkage visibleRule visible in code or documentationDomain intent only partially closed in current scope