diff --git a/home-claude/settings.json b/home-claude/settings.json
index a60b1da..ffdcc1f 100644
--- a/home-claude/settings.json
+++ b/home-claude/settings.json
@@ -53,5 +53,7 @@
   "enabledPlugins": {
     "csharp-lsp@claude-plugins-official": true
   },
-  "alwaysThinkingEnabled": true
+  "alwaysThinkingEnabled": true,
+  "effortLevel": "low",
+  "model": "sonnet"
 }
diff --git a/memory/projects/ISMS-DB/MEMORY.md b/memory/projects/ISMS-DB/MEMORY.md
index d345637..a2c7d11 100644
--- a/memory/projects/ISMS-DB/MEMORY.md
+++ b/memory/projects/ISMS-DB/MEMORY.md
@@ -24,6 +24,34 @@
 - Script validates: .lock, control exists, Id unique, Iteration unique
 - For the JSON heredoc: use python3 inline to generate UUIDs + timestamps, then echo
 
+## Dark Mode
+- Project DOES use `data-bs-theme="dark"` — always add dark-mode overrides for 3rd-party CSS (Toast UI, diff2html, etc.)
+- Pattern: `[data-bs-theme="dark"] .class { ... }` in `site.css`
+- Use `var(--bs-body-bg)`, `var(--bs-body-color)`, `var(--bs-tertiary-bg)`, `var(--bs-border-color)`
+
 ## Razor Gotcha
 - `RZ1031`: cannot put free-standing `@(expr)` in `<option>` attribute area
 - Fix: use `@if (cond) { <option selected> } else { <option> }` pattern instead
+
+## tenant_data_context.md
+- **MUST be kept up-to-date** when adding/changing entities, enums, storage patterns, or inheritance behavior
+- Located at repo root: `tenant_data_context.md`
+- Audience: AI agents working on tenant JSON data locally (after Fetch-Data.ps1 download)
+- Covers: entity schemas, enums, FK constraints, computed fields, agent rules, split storage, tenant inheritance
+- When adding a new entity/enum/file: update File Map, Entity Schema, Enum Reference, Folder Structure sections
+
+## viFlow HTML Import Pipeline
+- Scripts: `parse_viflow.py` → `vorlage/parsed_processes.json` → `generate_workflows.py`
+- Source: viFlow HTML export (e.g. `vorlage/chrka.html`, ~1.3MB)
+- `parse_viflow.py`: regex-based, splits by "Unterprozess:" sections, extracts 5-column step tables (Nr|Input|Title+Details|Output|Swimlane)
+- `generate_workflows.py`: creates Process entries, ProcessWorkflow+Actions, Resistec inherited items (English translations), idempotent (dedup by name)
+- HTML structure: `<p class="titleheading">Unterprozess:</p>` → `<p class="titlebig">Name</p>`, Stammdaten table, step rows with `processnumber`/`subprocesstitle`/`swimlane`/`remarksdetails`
+- Decision steps: name contains "?", need ≥2 transitions with labels (ja/nein)
+- Run with `PYTHONIOENCODING=utf-8` on Windows (Slovenian chars etc.)
+- Keep both scripts in repo root for re-use with future viFlow exports
+
+## Workflow Split Storage
+- Index: `process_workflows.json` (Actions=[], use ActionCount)
+- Individual: `workflows/{id}.json` (full object with Actions)
+- Auto-migration on first access if old single-file format detected
+- Pages use `wf.ActionCount` not `wf.Actions.Count` when reading from index
diff --git a/memory/projects/claude-meta/MEMORY.md b/memory/projects/claude-meta/MEMORY.md
new file mode 100644
index 0000000..4809d10
--- /dev/null
+++ b/memory/projects/claude-meta/MEMORY.md
@@ -0,0 +1,15 @@
+# Global Workflow Memory
+
+## Feature-Session Workflow
+
+When user says `Feature [topic]` at session start:
+1. Look for `docs/feature-[topic].md` in project root (case-insensitive match)
+2. If found: read it, confirm context loaded, summarize key points briefly
+3. If not found: inform user that no feature doc exists yet, suggest creating one
+
+When user says `aktualisiere deine Feature-Beschreibung` at session end:
+1. Locate the relevant `docs/feature-[topic].md`
+2. Update it with key insights, decisions, and findings from the current session
+3. Keep it structured: Goal, Context, Key Decisions, Open Questions, Progress
+
+File naming: `docs/feature-[featureName].md` (camelCase featureName, e.g. `feature-jobSystem.md`)
diff --git a/memory/projects/kbsV3/MEMORY.md b/memory/projects/kbsV3/MEMORY.md
index 6da2dd5..46937b4 100644
--- a/memory/projects/kbsV3/MEMORY.md
+++ b/memory/projects/kbsV3/MEMORY.md
@@ -1,7 +1,8 @@
 # KbsV3 Memory
 
 ## Architecture Guide
-- See [kbsv3-architecture.md](kbsv3-architecture.md) for: Job system, SAP RFC calls, gRPC communication, DB read/write patterns
+- See [kbsv3-architecture.md](kbsv3-architecture.md) for: SAP RFC calls, gRPC communication, DB read/write patterns
+- See [job-system.md](job-system.md) for: full job lifecycle, scheduler pattern, ILongRunningJobService vs IKbsJob, InputData/Payload flow
 
 ## Plotly.Blazor (v7.0.0)
 - `ModeFlag` namespace: `Plotly.Blazor.Traces.ScatterLib`
@@ -27,6 +28,7 @@
 - Do NOT add new `JobType` enum entries for `IKbsJob` workers — they all share `CallServiceHost`
 - `ILongRunningJobService` is for services like AD sync, Hydra — they have dedicated JobType enum entries
 
+
 ## HydraArchiveData
 - Entity: HydraArchiveReloadId, AcquisitionTime, PdatrefPparamDatasupId, DecimalValue
 - Covering index: (HydraArchiveReloadId, AcquisitionTime, PdatrefPparamDatasupId)
diff --git a/memory/projects/kbsV3/job-system.md b/memory/projects/kbsV3/job-system.md
new file mode 100644
index 0000000..3dc073c
--- /dev/null
+++ b/memory/projects/kbsV3/job-system.md
@@ -0,0 +1,90 @@
+# KbsV3 Job System – How It Works
+
+## Job Lifecycle (end-to-end)
+
+1. **Job created** – via UI (`/adsyncjob`), REST controller, or Kbs3SchedulerService
+   - Sets `JobDto.JobType`, `InputData` (JSON string), `JobSubtype`, `CreatedBy`
+   - Persisted to DB via `IJobDataService.AddJob()`
+2. **Job picked up** – background poller in Kbs3ServerService finds pending jobs
+3. **JobStartService.StartJob(job)** – resolves `JobRequirementFactoryService`, checks `CanStart()`
+4. **gRPC call** – server calls microservice via `IJobSubscribeService.Subscribe(JobDataRequest)`
+   - `JobDataRequest.InputData` = the JSON string from step 1
+5. **Worker execution** – microservice's `ILongRunningJobService.Run(request, ct)` or `IKbsJob` host
+6. **Streaming** – protocol updates stream back via gRPC `IAsyncEnumerable<JobProtocolUpdate>`
+7. **Persistence** – server saves `Protocol`/`ProtocolLine` entities
+
+## Creating a Job Programmatically
+
+### From UI page (direct DB)
+```csharp
+job.JobType = JobType.DeleteInactiveUsers;
+job.InputData = "{\"DryRun\":true}";  // raw JSON string
+job.JobSubtype = JobSubtypes.SyncZadGroup;
+await JobService.AddJob(job);  // IJobDataService
+```
+
+### From microservice / scheduler (via gRPC IApiComposer)
+```csharp
+await apiComposer.CreateKnownJob(new CreateJobRequest
+{
+    Jobtype = JobType.DeleteInactiveUsers.GetStringValue(),  // "DeleteInactiveUsers"
+    PayLoad = "{\"DryRun\":false}",  // string → JobDto.InputData as-is
+    Priority = 5
+});
+```
+`PayLoad` (string) is stored directly as `JobDto.InputData` in `CreateAndEnqueueJob` (no extra serialization).
+No `IKnownJob` definition required for this approach.
+
+## Multi-Service Jobs (e.g. DeleteInactiveUsers)
+- Multiple microservices (AD, Azure, DW, Helpdesk, SQL) each register a worker for the same `JobType`
+- All fire automatically when one job with that `JobType` is created – server broadcasts to all registered services
+- No special routing needed; just one `CreateKnownJob` call triggers all
+
+## ILongRunningJobService Workers
+- Dedicated `JobType` enum entry (e.g. `JobType.DeleteInactiveUsers`)
+- Implement `Run(JobDataRequest request, CancellationToken ct)`
+- Read input: `JsonConvert.DeserializeObject<T>(request.InputData)`
+- Auto-registered as keyed service by `JobType` string value in `IntegrationServiceBase.AutoInjectServices()`
+- Log via `protocolLogger.LogHeadline / LogInfo / LogWarn / LogError`
+
+## IKbsJob Workers (Kbs3SqlService pattern)
+- All share `JobType.CallServiceHost`
+- Auto-registered by class name (e.g. `"SyncOpexPurchaseOrdersService"`)
+- `IKnownJob.ServiceName` = `"modulename.ClassName"` → host resolves last segment to find worker
+- Do NOT add new `JobType` enum entries for these
+
+## Kbs3SchedulerService
+
+### Adding a new scheduled job
+1. Create `MyJob : BaseJob` in `Kbs3SchedulerService/Jobs/`
+   - Constructor: `(ILogger<MyJob> logger, IApiComposer apiComposer) : base(logger)`
+   - Implement `override string Name` and `override async Task Execute(CancellationToken, KeyValue[] settings)`
+   - Auto-registered via `AddAllTypesOf<IJob>()` – no manual DI
+2. Add entry to `appsettings.Development.json` (IsEnabled: false) and `appsettings.Production.json` (IsEnabled: true)
+   under `SchedulerSettings.Items`
+
+### Schedule patterns
+| Type | Meaning | Key fields |
+|------|---------|------------|
+| 1 | Daily | `Interval`, `StartTime` |
+| 2 | Weekly | `Interval`, `DaysOfWeek: ["Monday","Friday",...]` |
+| 3 | Monthly | `DaysOfMonth` |
+| 5 | Interval | `CustomInterval: "HH:mm:ss"` |
+
+- `StartTime` on the root item = time of day for execution
+- `Settings` array → read in `Execute` via `settings.FirstOrDefault(s => s.Key == "Foo")?.Value`
+
+### Internal flow
+`SchedulerHostedService` → `SeriesSchedulerManager` → `SeriesJobScheduler` (infinite loop)
+→ `_schedule.GetNextExecution(DateTime.Now)` → `Task.Delay(...)` → `job.ExecuteAsync(...)`
+
+## InputData / Payload flow
+```
+CreateJobRequest.PayLoad (string)
+  → ApiComposerService.CreateKnownJob()
+  → CreateJobService.CreateAndEnqueueJob()
+  → if (payload is string) input = payload as string   ← no extra serialization!
+  → JobDto.InputData
+  → JobDataRequest.InputData (sent to microservice via gRPC)
+  → worker reads request.InputData
+```