AP0.8: Operating Model Extension I (Attention, Blocker, Backlog, Milestone)
Deploy Development / deploy (push) Successful in 45s
Test Suite / pytest-backend (push) Successful in 58s
Test Suite / lint-backend (push) Successful in 2s
Test Suite / compose-smoke (push) Has been skipped
Test Suite / k6 /api/health Baseline (push) Successful in 18s
Test Suite / playwright-smoke (push) Successful in 12s
Deploy Development / deploy (push) Successful in 45s
Test Suite / pytest-backend (push) Successful in 58s
Test Suite / lint-backend (push) Successful in 2s
Test Suite / compose-smoke (push) Has been skipped
Test Suite / k6 /api/health Baseline (push) Successful in 18s
Test Suite / playwright-smoke (push) Successful in 12s
Schema 007, regelbasierte Attention/NextAction im Data Layer, CRUD fuer Blocker/Backlog/Meilensteine, Workspace-Widget und Initiative-Detail-Sektionen. 25 Capabilities. Tests fuer Remote-Pytest auf Pi angepasst (conftest Session-Guard). Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
1 parent
6dd9dd4f62
commit
fe3bc4f057
39 files changed
+4333
-22
No files matched your search
@@ -0,0 +1,808 @@
|
||||
# Kairo – System Target State
|
||||
## v0.1 – Zielzustand des angedachten Gesamtsystems
|
||||
|
||||
**Status:** konsolidiertes Zielbild / Nordstern-Dokument
|
||||
**Stand:** 2026-07-05
|
||||
**Zweck:** Beschreibung des angedachten späteren Zielzustands von Kairo als System, unabhängig von konkreten Implementierungsschnitten.
|
||||
|
||||
---
|
||||
|
||||
## 1. Executive Summary
|
||||
|
||||
Kairo ist langfristig keine einfache Aufgabenliste, keine reine Projektmanagement-App, kein Roadmap-Tool und keine freie Workflow-Automatisierungsplattform.
|
||||
|
||||
Kairo ist eine:
|
||||
|
||||
```text
|
||||
adaptive, methodengeführte Entwicklungs- und Steuerungsplattform
|
||||
```
|
||||
|
||||
Der Kern ist ein:
|
||||
|
||||
```text
|
||||
Method-driven Adaptive Steering Core
|
||||
```
|
||||
|
||||
Kairo soll unterschiedliche Entwicklungs- und Steuerungskontexte führen können:
|
||||
|
||||
```text
|
||||
- Software- und Produktentwicklung
|
||||
- Buch- und Content-Entwicklung
|
||||
- Programmsteuerung
|
||||
- Projekt- und Initiativensteuerung
|
||||
- persönliche Entwicklung
|
||||
- Reifegradentwicklung
|
||||
- Lernpfade
|
||||
- Routinen / Betrieb
|
||||
- Review- und Auditkontexte
|
||||
- Agentenläufe
|
||||
- KI-/Tool-gestützte Steuerung
|
||||
```
|
||||
|
||||
Die zentrale Leitfrage lautet:
|
||||
|
||||
```text
|
||||
Welcher nächste Schritt bringt diesen steuerbaren Kontext jetzt am wirksamsten voran?
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 2. Grundverständnis
|
||||
|
||||
Kairo steuert nicht nur Aufgaben.
|
||||
|
||||
Kairo steuert Entwicklungen.
|
||||
|
||||
Eine Entwicklung kann sein:
|
||||
|
||||
```text
|
||||
- ein Produkt
|
||||
- ein Programm
|
||||
- ein Projekt
|
||||
- eine Initiative
|
||||
- ein Buch
|
||||
- eine Fähigkeit
|
||||
- ein Reifegradziel
|
||||
- eine Routine
|
||||
- ein Review
|
||||
- ein Audit
|
||||
- ein Agentenauftrag
|
||||
```
|
||||
|
||||
Kairo hält dabei fest:
|
||||
|
||||
```text
|
||||
- Was soll entwickelt werden?
|
||||
- Warum ist es relevant?
|
||||
- Welche Methode gilt?
|
||||
- Welche Struktur braucht diese Methode?
|
||||
- Wo steht der Kontext aktuell?
|
||||
- Was fehlt?
|
||||
- Was blockiert?
|
||||
- Was braucht Aufmerksamkeit?
|
||||
- Was ist der nächste wirksame Schritt?
|
||||
- Wer oder welcher Agent soll handeln?
|
||||
- Worauf wartet Kairo?
|
||||
- Welche Nachweise, Reviews oder Entscheidungen sind nötig?
|
||||
- Wann wird angepasst, eskaliert oder abgeschlossen?
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 3. Produktidentität
|
||||
|
||||
Kairo ist:
|
||||
|
||||
```text
|
||||
System of Record
|
||||
Steering Authority
|
||||
Actor-/Tenant-sicherer Kontextgeber
|
||||
methodengeführter Steuerungskern
|
||||
Roadmap- und Entwicklungsstrukturmodell
|
||||
Signal- und NextAction-System
|
||||
später Workflow-/Agenten-/Tool-orchestrierend
|
||||
```
|
||||
|
||||
Kairo ist nicht:
|
||||
|
||||
```text
|
||||
freie Workflow-Plattform
|
||||
beliebiger Automatisierungsbaukasten
|
||||
reines Projektmanagementsystem
|
||||
reine To-do-Liste
|
||||
reines Roadmap-Tool
|
||||
reine KI-Agentenplattform
|
||||
reines Prompt-System
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 4. Kernmodell
|
||||
|
||||
Das Zielmodell:
|
||||
|
||||
```text
|
||||
Steerable Object
|
||||
→ Steering Context
|
||||
→ Steering Method
|
||||
→ Standard Lifecycle / Method Lifecycle
|
||||
→ Hook Slugs
|
||||
→ Structure Builder
|
||||
→ Signal Rules
|
||||
→ NextAction Strategy
|
||||
→ Assignment Strategy
|
||||
→ Waiting / Reminder / Escalation Strategy
|
||||
→ Result Intake
|
||||
→ Evidence / Review / Decision
|
||||
→ Adaptation / Closure
|
||||
→ optionale Workflow-Fragmente
|
||||
→ optionale Human / Agent / LLM / Tool Steps
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 5. Method-driven Adaptive Steering Core
|
||||
|
||||
Der Adaptive Steering Core ist der fachliche und technische Kern von Kairo.
|
||||
|
||||
Er führt steuerbare Kontexte durch methodengeführte Steuerungsabläufe.
|
||||
|
||||
Typischer Ablauf:
|
||||
|
||||
```text
|
||||
1. Ziel / Entwicklungsabsicht erfassen
|
||||
2. Name, Scope und Kontext definieren
|
||||
3. passende Basismethode wählen oder vorschlagen
|
||||
4. gemäß Methode eine Steuerungsstruktur erzeugen
|
||||
- Roadmap
|
||||
- Backlog
|
||||
- WBS
|
||||
- Reifegradpfad
|
||||
- Kapitelstruktur
|
||||
- Feature Landscape
|
||||
- WorkCycles
|
||||
- Review Gates
|
||||
5. Next Best Action ableiten
|
||||
6. Actor zuweisen
|
||||
7. auf Ergebnis, Rückmeldung, Termin oder Ereignis warten
|
||||
8. erinnern, eskalieren oder nachsteuern
|
||||
9. Ergebnis zurückführen
|
||||
10. prüfen, reviewen, adaptieren oder abschließen
|
||||
```
|
||||
|
||||
Dieser Ablauf ist der methodengeführte Steuerungsworkflow.
|
||||
|
||||
Er ist nicht gleichbedeutend mit einer freien technischen Workflow Engine.
|
||||
|
||||
---
|
||||
|
||||
## 6. Standard Lifecycle
|
||||
|
||||
Kairo besitzt einen minimalen Standard Lifecycle als gemeinsame Basis.
|
||||
|
||||
```text
|
||||
intake
|
||||
method_selection
|
||||
structure_setup
|
||||
planning
|
||||
action_selection
|
||||
assignment
|
||||
waiting
|
||||
result_intake
|
||||
validation
|
||||
review
|
||||
adaptation
|
||||
closure
|
||||
```
|
||||
|
||||
Methoden können diesen Lifecycle:
|
||||
|
||||
```text
|
||||
- verwenden
|
||||
- teilweise überspringen
|
||||
- spezialisieren
|
||||
- erweitern
|
||||
- eigene Hooks ergänzen
|
||||
```
|
||||
|
||||
Der Standard Lifecycle ist kein starrer Projektprozess, sondern ein gemeinsames Steuerungsraster.
|
||||
|
||||
---
|
||||
|
||||
## 7. Steering Methods
|
||||
|
||||
Eine Methode ist kein Label, sondern ein registriertes Steuerungspaket.
|
||||
|
||||
Eine Methode definiert:
|
||||
|
||||
```text
|
||||
- supported_domains
|
||||
- supported_scope_types
|
||||
- default_lifecycle
|
||||
- required_structures
|
||||
- allowed_structures
|
||||
- hook_slugs
|
||||
- structure_builders
|
||||
- attention_rule_providers
|
||||
- next_action_strategy
|
||||
- assignment_strategy
|
||||
- waiting_strategy
|
||||
- reminder_strategy
|
||||
- escalation_strategy
|
||||
- result_intake_strategy
|
||||
- review_strategy
|
||||
- evidence_policy
|
||||
- closure_policy
|
||||
- allowed_workflow_fragments
|
||||
- allowed_node_types
|
||||
- default_configuration
|
||||
```
|
||||
|
||||
Methoden werden zunächst built-in und code-first angelegt, später versioniert und begrenzt konfigurierbar.
|
||||
|
||||
Langfristig können Method Profiles entstehen.
|
||||
|
||||
---
|
||||
|
||||
## 8. Method Registry
|
||||
|
||||
Kairo benötigt eine Method Registry.
|
||||
|
||||
Sie funktioniert ähnlich einem Widget-System:
|
||||
|
||||
```text
|
||||
Eine Methode meldet sich mit Slug, Version, unterstützten Domains,
|
||||
Lifecycle, Hooks, Structure Builders, Strategies und Policies an.
|
||||
```
|
||||
|
||||
Ziel:
|
||||
|
||||
```text
|
||||
Neue Methoden sollen ergänzt werden können,
|
||||
ohne den Core jedes Mal umzubauen.
|
||||
```
|
||||
|
||||
Beispiele für Methoden:
|
||||
|
||||
```text
|
||||
simple_action_driven
|
||||
product_milestone_driven
|
||||
backlog_driven
|
||||
wbs_driven
|
||||
cycle_based
|
||||
content_chapter_based
|
||||
maturity_progression
|
||||
routine_control
|
||||
review_driven
|
||||
audit_driven
|
||||
agent_controlled
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 9. Steering Domains
|
||||
|
||||
Kairo unterstützt unterschiedliche Steuerungsdomänen.
|
||||
|
||||
Beispiele:
|
||||
|
||||
```text
|
||||
program
|
||||
project
|
||||
product_development
|
||||
content_development
|
||||
personal_development
|
||||
maturity_development
|
||||
routine_operations
|
||||
audit_review
|
||||
agent_execution
|
||||
custom
|
||||
```
|
||||
|
||||
Jede Domain kann eigene typische Lifecycles, Roadmap-Strukturen, Methoden, Hooks und Signals haben.
|
||||
|
||||
---
|
||||
|
||||
## 10. Steerable Objects
|
||||
|
||||
Steuerbar sind nicht nur Initiativen.
|
||||
|
||||
Mögliche steuerbare Objekte:
|
||||
|
||||
```text
|
||||
program
|
||||
initiative
|
||||
project
|
||||
roadmap
|
||||
roadmap_item
|
||||
backlog_item
|
||||
action
|
||||
blocker
|
||||
review
|
||||
evidence
|
||||
decision
|
||||
recurring_element
|
||||
work_cycle
|
||||
agent_run
|
||||
custom
|
||||
```
|
||||
|
||||
Diese Objekte werden über Steering Contexts an Methoden und Domains gebunden.
|
||||
|
||||
---
|
||||
|
||||
## 11. Steering Context
|
||||
|
||||
Ein Steering Context beschreibt:
|
||||
|
||||
```text
|
||||
Für welches steuerbare Objekt gilt welche Methode, Domain und Lifecycle-Logik?
|
||||
```
|
||||
|
||||
Konzeptuell:
|
||||
|
||||
```text
|
||||
SteeringContext
|
||||
- tenant_id
|
||||
- scope_type
|
||||
- scope_id
|
||||
- steering_domain
|
||||
- lifecycle_model
|
||||
- steering_method
|
||||
- method_profile_id optional
|
||||
- method_version
|
||||
- configuration_json
|
||||
- inherits_from_parent
|
||||
- is_active
|
||||
```
|
||||
|
||||
Untergeordnete Objekte können den Steering Context ihres Parents erben oder überschreiben.
|
||||
|
||||
---
|
||||
|
||||
## 12. Roadmap / Development Model
|
||||
|
||||
Roadmap ist ein Kernmodell, nicht nur eine UI-Ansicht.
|
||||
|
||||
Kairo braucht Roadmaps, um Entwicklungsrichtung und Entwicklungsstruktur abzubilden.
|
||||
|
||||
Roadmap-Struktur:
|
||||
|
||||
```text
|
||||
Roadmap
|
||||
→ Roadmap Lane
|
||||
→ Roadmap Item
|
||||
```
|
||||
|
||||
RoadmapItems können sein:
|
||||
|
||||
```text
|
||||
milestone
|
||||
maturity_stage
|
||||
feature
|
||||
capability
|
||||
phase
|
||||
release
|
||||
chapter
|
||||
research_topic
|
||||
work_cycle
|
||||
learning_step
|
||||
review_gate
|
||||
decision_gate
|
||||
architecture_step
|
||||
dependency
|
||||
risk_reduction
|
||||
custom
|
||||
```
|
||||
|
||||
Entscheidung:
|
||||
|
||||
```text
|
||||
Milestone = RoadmapItem(type = milestone)
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 13. Roadmap vs. Backlog vs. Action
|
||||
|
||||
Zentrale Unterscheidung:
|
||||
|
||||
```text
|
||||
RoadmapItem = Entwicklungsziel / Orientierung / geplanter Entwicklungsschritt
|
||||
BacklogItem = konkreter möglicher Handlungsbedarf
|
||||
Action = freigegebene operative Maßnahme
|
||||
```
|
||||
|
||||
Beispiel Softwareentwicklung:
|
||||
|
||||
```text
|
||||
RoadmapItem:
|
||||
Attention Engine regelbasiert
|
||||
|
||||
BacklogItems:
|
||||
- Attention DTO definieren
|
||||
- Blockierte Maßnahmen auswerten
|
||||
- Widget bauen
|
||||
|
||||
Actions:
|
||||
- Endpoint implementieren
|
||||
- Tests schreiben
|
||||
```
|
||||
|
||||
Beispiel Buchentwicklung:
|
||||
|
||||
```text
|
||||
RoadmapItem:
|
||||
Kapitel 3 – Argumentation ausarbeiten
|
||||
|
||||
BacklogItems:
|
||||
- Recherche zur These prüfen
|
||||
- Beispiel suchen
|
||||
- Rohfassung schreiben
|
||||
|
||||
Actions:
|
||||
- Heute Abschnitt 3.1 entwerfen
|
||||
```
|
||||
|
||||
Beispiel Reifegradentwicklung:
|
||||
|
||||
```text
|
||||
RoadmapItem:
|
||||
Reifegrad: stabile Seitkick-Beweglichkeit
|
||||
|
||||
BacklogItems:
|
||||
- Hüftmobilität prüfen
|
||||
- Dehnroutine auswählen
|
||||
- Technikvideo aufnehmen
|
||||
|
||||
Actions:
|
||||
- 10 Minuten Mobility durchführen
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 14. Structure Builder
|
||||
|
||||
Structure Builder erzeugen methodenspezifische Steuerungsstrukturen.
|
||||
|
||||
Beispiele:
|
||||
|
||||
```text
|
||||
roadmap_builder
|
||||
milestone_plan_builder
|
||||
feature_landscape_builder
|
||||
wbs_builder
|
||||
backlog_builder
|
||||
maturity_path_builder
|
||||
chapter_structure_builder
|
||||
cycle_structure_builder
|
||||
review_gate_builder
|
||||
routine_structure_builder
|
||||
```
|
||||
|
||||
Sie sind zentrale Erweiterungspunkte für Methoden.
|
||||
|
||||
Ein Structure Builder legt z. B. an:
|
||||
|
||||
```text
|
||||
- Roadmap
|
||||
- Roadmap Lanes
|
||||
- Roadmap Items
|
||||
- initiale BacklogItems
|
||||
- Review Gates
|
||||
- WorkCycles
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 15. Hooks
|
||||
|
||||
Hooks sind stabile Einhängepunkte in Steuerungsabläufen.
|
||||
|
||||
Beispiele:
|
||||
|
||||
```text
|
||||
on_goal_captured
|
||||
on_scope_named
|
||||
on_method_selection_required
|
||||
on_method_selected
|
||||
on_structure_required
|
||||
on_roadmap_required
|
||||
on_backlog_required
|
||||
on_wbs_required
|
||||
on_maturity_path_required
|
||||
on_content_structure_required
|
||||
on_plan_required
|
||||
on_dependency_analysis_required
|
||||
on_dod_definition_required
|
||||
on_next_action_requested
|
||||
on_assignment_required
|
||||
on_wait_started
|
||||
on_due_date_reached
|
||||
on_result_overdue
|
||||
on_reminder_required
|
||||
on_escalation_required
|
||||
on_result_received
|
||||
on_evidence_required
|
||||
on_review_due
|
||||
on_reassessment_required
|
||||
on_replan_required
|
||||
on_closure_requested
|
||||
```
|
||||
|
||||
An Hooks können später kleine Workflow-Fragmente, Rules, Prompts, Agent Steps oder Tool Steps gebunden werden.
|
||||
|
||||
Hooks sind nicht beliebige technische Events, sondern fachlich stabile Steuerungspunkte.
|
||||
|
||||
---
|
||||
|
||||
## 16. Strategies
|
||||
|
||||
Methoden bringen oder referenzieren Strategies.
|
||||
|
||||
Zentrale Strategy-Typen:
|
||||
|
||||
```text
|
||||
NextActionStrategy
|
||||
AttentionRuleProvider
|
||||
AssignmentStrategy
|
||||
WaitingStrategy
|
||||
ReminderStrategy
|
||||
EscalationStrategy
|
||||
ResultIntakeStrategy
|
||||
EvidencePolicy
|
||||
ReviewStrategy
|
||||
ClosurePolicy
|
||||
```
|
||||
|
||||
Diese verhindern, dass methodenspezifische Logik monolithisch im Core wächst.
|
||||
|
||||
---
|
||||
|
||||
## 17. Signals
|
||||
|
||||
Signals sind sichtbare Ausgaben des Steuerungskerns.
|
||||
|
||||
Signal Types:
|
||||
|
||||
```text
|
||||
attention
|
||||
next_action
|
||||
review_due
|
||||
evidence_required
|
||||
escalation
|
||||
blocker_signal
|
||||
progress_signal
|
||||
risk_signal
|
||||
decision_required
|
||||
dependency_signal
|
||||
dod_signal
|
||||
workflow_trigger
|
||||
```
|
||||
|
||||
### Attention
|
||||
|
||||
Beantwortet:
|
||||
|
||||
```text
|
||||
Was braucht Aufmerksamkeit?
|
||||
```
|
||||
|
||||
### NextActionCandidate
|
||||
|
||||
Beantwortet:
|
||||
|
||||
```text
|
||||
Was ist der nächste wirksame Schritt?
|
||||
```
|
||||
|
||||
Signals müssen erklärbar sein:
|
||||
|
||||
```text
|
||||
- warum erzeugt?
|
||||
- durch welche Methode?
|
||||
- durch welchen Hook?
|
||||
- auf welchen Scope bezogen?
|
||||
- welche Datenquelle?
|
||||
- welche empfohlene Handlung?
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 18. Assignment / Waiting / Reminder / Escalation
|
||||
|
||||
Der Steuerungskern endet nicht mit dem Erzeugen einer Action.
|
||||
|
||||
Kairo muss steuern:
|
||||
|
||||
```text
|
||||
- Actor auswählen
|
||||
- Actor zuweisen
|
||||
- auf Ergebnis warten
|
||||
- Fälligkeit überwachen
|
||||
- erinnern
|
||||
- eskalieren
|
||||
- externe Ereignisse aufnehmen
|
||||
- Agentenergebnisse zurückführen
|
||||
```
|
||||
|
||||
Actors können sein:
|
||||
|
||||
```text
|
||||
human
|
||||
agent
|
||||
working_group
|
||||
external_system
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 19. Result Intake / Evidence / Review / Decision
|
||||
|
||||
Kairo führt Ergebnisse zurück und entscheidet, was daraus folgt.
|
||||
|
||||
Mögliche Rückführungen:
|
||||
|
||||
```text
|
||||
- Action erledigt
|
||||
- Ergebnis unvollständig
|
||||
- Evidence erforderlich
|
||||
- Review erforderlich
|
||||
- Decision erforderlich
|
||||
- Blocker entstanden
|
||||
- RoadmapItem erreicht
|
||||
- Reifegrad plausibel erhöht
|
||||
- neue NextAction nötig
|
||||
```
|
||||
|
||||
Evidence, Reviews und Decisions sind zentrale Qualitätssicherungs- und Steuerungsobjekte.
|
||||
|
||||
---
|
||||
|
||||
## 20. Workflow-Fragmente und Runtime
|
||||
|
||||
Kairo soll langfristig Workflow-Fähigkeit haben.
|
||||
|
||||
Aber:
|
||||
|
||||
```text
|
||||
Workflows sind methodengeführte Steuerungsabläufe,
|
||||
keine beliebige freie Automatisierungsplattform als Produktkern.
|
||||
```
|
||||
|
||||
Zielmodell:
|
||||
|
||||
```text
|
||||
Hook
|
||||
→ Workflow Fragment
|
||||
→ Node / Step
|
||||
→ Result
|
||||
→ Rückführung in Kairo
|
||||
```
|
||||
|
||||
Workflow-Fragmente können später enthalten:
|
||||
|
||||
```text
|
||||
human_task
|
||||
approval
|
||||
decision_gate
|
||||
review
|
||||
reflection
|
||||
llm_prompt
|
||||
agent_task
|
||||
tool_call
|
||||
state_transition
|
||||
evidence_check
|
||||
dod_check
|
||||
dependency_analysis
|
||||
notification
|
||||
reminder
|
||||
escalation
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 21. Agenten / LLMs / Tools
|
||||
|
||||
Kairo bleibt Steering Authority.
|
||||
|
||||
Agenten, LLMs und Tools sind Provider.
|
||||
|
||||
Sie können später an Hooks und Workflow-Fragmente angebunden werden.
|
||||
|
||||
Grundsatz:
|
||||
|
||||
```text
|
||||
Kairo entscheidet Kontext, Methode, Auftrag und Rückführung.
|
||||
Agenten / LLMs / Tools führen einzelne Steps aus.
|
||||
```
|
||||
|
||||
Agenten sind Actors.
|
||||
|
||||
Das ist wichtig für:
|
||||
|
||||
```text
|
||||
- Verantwortlichkeit
|
||||
- Audit
|
||||
- Assignments
|
||||
- Sichtbarkeit
|
||||
- Tenant-Kontext
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 22. Security / Governance / Audit
|
||||
|
||||
Kairo muss tenant- und actor-sicher bleiben.
|
||||
|
||||
Grundsätze:
|
||||
|
||||
```text
|
||||
- tenant_id immer aus TenantContext
|
||||
- keine tenant_id aus Client Input vertrauen
|
||||
- Actor-Bezüge tenant-intern validieren
|
||||
- Capabilities prüfen
|
||||
- Router dünn
|
||||
- Services schreiben
|
||||
- Data Layer liest
|
||||
- Audit-by-design
|
||||
- Signals und NextActions erklärbar
|
||||
- Agentenaktionen nachvollziehbar
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 23. UI-Zielbild
|
||||
|
||||
Die UI dient als Steuerungs- und Interaktionsschicht.
|
||||
|
||||
Langfristige Views:
|
||||
|
||||
```text
|
||||
Workspace
|
||||
Steering View
|
||||
Roadmap View
|
||||
Method View
|
||||
Attention / NextAction View
|
||||
Backlog View
|
||||
Action View
|
||||
Waiting / Reminder View
|
||||
Evidence / Review / Decision View
|
||||
Workflow Run View später
|
||||
Method/Profile Admin später
|
||||
Agent Run View später
|
||||
```
|
||||
|
||||
Die UI darf nicht die Methode hart kodieren.
|
||||
|
||||
Sie muss methodenbewusst und strukturgetrieben sein.
|
||||
|
||||
---
|
||||
|
||||
## 24. Architektur-Evolutionslinie
|
||||
|
||||
Grobe Evolutionslinie:
|
||||
|
||||
```text
|
||||
1. Existing Foundation
|
||||
2. Adaptive Steering Core Foundation
|
||||
3. Method Registry / Hook Registry / Standard Lifecycle
|
||||
4. Roadmap / Structure Builder Foundation
|
||||
5. Signal / NextAction Engine
|
||||
6. Assignment / Waiting / Reminder / Escalation
|
||||
7. Evidence / Review / Decision
|
||||
8. Workflow Fragments and Bindings
|
||||
9. Workflow Runtime
|
||||
10. Agent / LLM / Tool Providers
|
||||
11. Method Profiles
|
||||
12. Method Designer / Advanced Configuration
|
||||
13. Advanced AI Steering / Next Best Action Ranking
|
||||
```
|
||||
|
||||
Dies ist keine Sprintplanung, sondern ein Zielbild der technischen Entwicklung.
|
||||
|
||||
---
|
||||
|
||||
## 25. Zielzustand in einem Satz
|
||||
|
||||
Kairo soll ein methodengeführter adaptiver Steuerungskern werden, der Entwicklungen unterschiedlichster Art über registrierbare Methoden, gemeinsame Kernstrukturen, Roadmaps, Hooks, Signals, NextActions, Assignments, Waiting-/Reminder-/Escalation-Mechanismen und später Agent-/Workflow-Fragmente wirksam und nachvollziehbar steuert.
|
||||
@@ -1,7 +1,7 @@
|
||||
# AP0.8 – Operating Model Extension I
|
||||
## Implementierungsauftrag zur Prüfung v0.1
|
||||
|
||||
**Status:** Entwurf zur Freigabe
|
||||
**Status:** ersetzt durch `Sprint0_AP0_8_Assignment_v0.2.md`
|
||||
**Stand:** 2026-07-05
|
||||
**Vorgänger:** AP0.7 ✓ · AP0.R1 ✓
|
||||
**Zielversion:** `0.8.0-ap0.8` · Schema `007`
|
||||
|
||||
@@ -0,0 +1,402 @@
|
||||
# AP0.8 – Operating Model Extension I
|
||||
## Implementierungsauftrag v0.2 (freigegeben)
|
||||
|
||||
**Status:** freigegeben zur Umsetzung
|
||||
**Stand:** 2026-07-05
|
||||
**Vorgänger:** AP0.7 ✓ · AP0.R1 ✓ · KAIRO-ARCH-01 ✓
|
||||
**Zielversion:** `0.8.0-ap0.8` · Schema `007`
|
||||
**Ersetzt:** `Sprint0_AP0_8_Assignment_v0.1.md` (Entwurf)
|
||||
|
||||
---
|
||||
|
||||
## Einordnung
|
||||
|
||||
AP0.8 ist der **erste fachliche Ausbau nach dem Product Reset** und der **erste Schritt der Operating-Model-Evolutionslinie** gemäß System Target State §24:
|
||||
|
||||
```text
|
||||
1. Existing Foundation ✓ AP0.1–AP0.7
|
||||
2. Adaptive Steering Core — nicht AP0.8 (→ AP1.x)
|
||||
3. Method Registry / Hooks — nicht AP0.8
|
||||
4. Roadmap / Structure Builder — nicht AP0.8 (Milestone als MVP-Brücke)
|
||||
5. Signal / NextAction Engine ← AP0.8a startet hier (regelbasiert, Data Layer)
|
||||
6. Assignment / Waiting / … — teilweise vorhanden; Waiting → AP0.9+
|
||||
7. Evidence / Review / Decision — AP0.9
|
||||
```
|
||||
|
||||
Ziel: Kairo sichtbar über `Initiative → Action` hinausheben und die Leitfrage operationalisieren:
|
||||
|
||||
> Welcher nächste Schritt bringt ein Vorhaben aktuell am wirkungsvollsten voran?
|
||||
|
||||
AP0.8 baut auf AP0.7 auf (Tenant-Invarianten, Actor Directory, Data Layer) und bleibt **regelbasiert ohne KI, ohne Method Registry, ohne SteeringContext**.
|
||||
|
||||
---
|
||||
|
||||
## Verbindliche Referenzen
|
||||
|
||||
Vor Umsetzung lesen:
|
||||
|
||||
1. `docs/product/Kairo_Product_Definition_and_MVP_Reset_v0.1.md`
|
||||
2. `docs/product/Kairo_Canonical_Operating_Model_v0.1.md` — §5 Objektmodell, §7 Attention Logic, §8 Data Layer
|
||||
3. `docs/product/Kairo_Corrected_MVP_Roadmap_v0.1.md` — AP0.8
|
||||
4. `docs/architecture/Kairo_System_Target_State_v0.1.md` — §13 Roadmap/Backlog/Action, §17 Signals
|
||||
5. `docs/architecture/Kairo_Target_Architecture_Method_Driven_Adaptive_Steering_Core_v0.1.md` — Phasen A/B, AD-TA-08
|
||||
6. `docs/architecture/Kairo_Tenant_Invariants_v0.1.md`
|
||||
7. `docs/sprints/Sprint0_AP0_7_Completion_Report_v0.2.md`
|
||||
|
||||
---
|
||||
|
||||
## Abgleich System Target State
|
||||
|
||||
| Zielbild (System Target State) | AP0.8-Umsetzung |
|
||||
|-------------------------------|-----------------|
|
||||
| RoadmapItem = Entwicklungsziel; BacklogItem = möglicher Handlungsbedarf; Action = committete Maßnahme | Backlog und Action **getrennt**; Convert-Flow |
|
||||
| Milestone = RoadmapItem(type=milestone) langfristig | **MVP-Brücke:** eigene `milestones`-Tabelle; Konsolidierung → RoadmapItem in Phase C (AD-TA-08) |
|
||||
| Attention + NextActionCandidate als Signal-Typen | Read Models in `data_layer/attention.py` |
|
||||
| Signals erklärbar (reason, scope, data_source) | DTO-Felder `reason_code`, `scope_type`, `scope_id`, `data_source` |
|
||||
| Blocker als Steuerungsobjekt | eigene Tabelle, nicht nur Action-Status |
|
||||
| Kein SteeringContext / Method Registry yet | explizit Nicht-Scope |
|
||||
| Kein Workflow / KI | explizit Nicht-Scope |
|
||||
|
||||
---
|
||||
|
||||
## Leitentscheidungen (freigegeben)
|
||||
|
||||
| Entscheidung | Entscheidung | Begründung |
|
||||
|--------------|--------------|------------|
|
||||
| Umsetzungsreihenfolge | 4 Slices (8a→8d) in **einem** AP0.8 | Attention zuerst für Produktwirkung |
|
||||
| Blocker | **Eigenes Objekt** | Canonical Model + System Target State §10 |
|
||||
| Backlog vs. Action | **Getrennte Tabellen** | §13 Unterscheidung Program Director vs. To-do |
|
||||
| Milestone | **Eigene Tabelle (MVP-Brücke)** | Schnell prüfbar; Roadmap-Konsolidierung später |
|
||||
| Project | **Nicht in AP0.8** | Milestone direkt an Initiative |
|
||||
| Action-Status `ready`/`review_required` | **Nicht in AP0.8** | AP0.9 |
|
||||
| Due Dates auf Actions | **Nicht in AP0.8** | Regel 7 stub; Milestone `target_date` optional |
|
||||
| Attention: open Blockers | **Ja, Regel 1b** | Ergänzung zu blocked Actions |
|
||||
| Capabilities | 7 neue (1 read + 3×2) | Registry-first |
|
||||
| `kairo.attention.read` | **Eigene Capability** | Explizit, Grants wie workspace.read |
|
||||
| Member manage Blocker/Backlog/Milestone | **Ja** | wie Actions |
|
||||
| Convert Backlog→Action | **Pflicht, minimal** | Commit-Flow |
|
||||
| Blocker ↔ action.status=blocked | **Optional bei Create** | nicht zwingend synchron |
|
||||
| UI | AP0.6b-Optik | kein Redesign |
|
||||
|
||||
---
|
||||
|
||||
## Empfohlene Umsetzungsreihenfolge
|
||||
|
||||
```text
|
||||
AP0.8a Attention / NextActionCandidate (Read Model + Widget)
|
||||
AP0.8b Blocker (Domäne + API + UI)
|
||||
AP0.8c BacklogItem (Domäne + API + UI + convert)
|
||||
AP0.8d Milestone minimal (Domäne + API + UI + Attention Regel 6)
|
||||
```
|
||||
|
||||
Reihenfolge ist verbindlich für die Implementierung.
|
||||
|
||||
---
|
||||
|
||||
# Teil A — AP0.8a: Attention / NextActionCandidate
|
||||
|
||||
## Ziel
|
||||
|
||||
Regelbasierte **Signal-Schicht** im Data Layer — erste Instanz der Signal/NextAction-Engine (System Target State §17). Später Refactor in `backend/steering/signals/` mit Method Rule Providers.
|
||||
|
||||
## Data Layer
|
||||
|
||||
Neue Datei: `backend/data_layer/attention.py`
|
||||
|
||||
```text
|
||||
get_attention_items(ctx) -> list[AttentionItemDTO]
|
||||
get_next_action_candidates(ctx, *, limit=10) -> list[NextActionCandidateDTO]
|
||||
```
|
||||
|
||||
**AttentionItemDTO** (mit Erklärbarkeit):
|
||||
|
||||
```json
|
||||
{
|
||||
"kind": "blocked_action | open_blocker | high_priority_action | unassigned_action | initiative_without_next_action | stale_initiative | milestone_at_risk",
|
||||
"severity": "info | warning | critical",
|
||||
"title": "...",
|
||||
"summary": "...",
|
||||
"scope_type": "action | blocker | initiative | milestone",
|
||||
"scope_id": "...",
|
||||
"initiative_id": "...",
|
||||
"action_id": "...",
|
||||
"blocker_id": "...",
|
||||
"milestone_id": "...",
|
||||
"reason_code": "...",
|
||||
"data_source": "actions | blockers | initiatives | milestones | backlog_items"
|
||||
}
|
||||
```
|
||||
|
||||
**NextActionCandidateDTO:**
|
||||
|
||||
```json
|
||||
{
|
||||
"kind": "assign_action | resolve_blocker | create_action | convert_backlog | review_milestone",
|
||||
"title": "...",
|
||||
"summary": "...",
|
||||
"initiative_id": "...",
|
||||
"action_id": "...",
|
||||
"backlog_item_id": "...",
|
||||
"reason_code": "...",
|
||||
"recommended_action": "..."
|
||||
}
|
||||
```
|
||||
|
||||
### Regeln
|
||||
|
||||
| # | Regel | Datenquelle | Severity |
|
||||
|---|-------|-------------|----------|
|
||||
| 1 | Blockierte Maßnahmen | `actions.status = blocked` | critical |
|
||||
| 1b | Offene Blocker | `blockers.status IN (open, in_progress)` | warning |
|
||||
| 2 | High-Priority offen | `priority = high`, open/in_progress; **personal** wenn Actor zugewiesen | warning |
|
||||
| 3 | Maßnahmen ohne Assignment | keine `action_assignments` | warning |
|
||||
| 4 | Vorhaben ohne offene Maßnahme | active/paused, keine open/in_progress/blocked Action | info |
|
||||
| 5 | Lange unveränderte Vorhaben | `updated_at` > 14 Tage, status active | info |
|
||||
| 6 | Meilenstein at_risk | ab AP0.8d: `milestones.status = at_risk` | warning |
|
||||
| 7–9 | Due Dates, Reviews, Recurring | AP0.9+ — stub/leer | — |
|
||||
|
||||
**NextAction-Kandidaten (regelbasiert, limitiert):**
|
||||
|
||||
- unassigned Action → `assign_action`
|
||||
- blocked Action / open Blocker → `resolve_blocker`
|
||||
- Initiative ohne offene Maßnahme → `create_action`
|
||||
- accepted BacklogItem ohne Action → `convert_backlog`
|
||||
|
||||
Sortierung Attention: critical → warning → info. Keine Client-Priorisierung.
|
||||
|
||||
## API
|
||||
|
||||
```text
|
||||
GET /api/workspace/attention
|
||||
GET /api/workspace/next-actions?limit=10
|
||||
```
|
||||
|
||||
Capability: **`kairo.attention.read`**
|
||||
|
||||
## Frontend
|
||||
|
||||
- **`AttentionWidget`** (Registry-Key `kairo.attention`, order 8, unter Überblick)
|
||||
- Links zu Vorhaben/Maßnahmen/Blockern
|
||||
- Loading / Error / Empty States
|
||||
|
||||
## Tests
|
||||
|
||||
- Regeln 1, 1b, 2–5; Regel 6 nach 8d
|
||||
- NextAction: convert_backlog, create_action
|
||||
- Cross-Tenant isolation
|
||||
- Router delegiert an Data Layer
|
||||
|
||||
---
|
||||
|
||||
# Teil B — AP0.8b: Blocker
|
||||
|
||||
## Migration `007_operating_model_extension_i.sql`
|
||||
|
||||
```sql
|
||||
blockers (
|
||||
id UUID PK,
|
||||
tenant_id UUID NOT NULL REFERENCES tenants(id),
|
||||
initiative_id UUID NOT NULL REFERENCES initiatives(id),
|
||||
action_id UUID NULL REFERENCES actions(id),
|
||||
title VARCHAR(255) NOT NULL,
|
||||
description TEXT DEFAULT '',
|
||||
status VARCHAR(32) NOT NULL, -- open, in_progress, resolved, accepted_risk, dismissed
|
||||
reported_by_actor_id UUID NULL REFERENCES actors(id),
|
||||
created_at, updated_at
|
||||
)
|
||||
```
|
||||
|
||||
Index: `(tenant_id, initiative_id)`, `(tenant_id, status)`
|
||||
|
||||
## Service + API
|
||||
|
||||
`backend/services/blockers.py` — CRUD, Audit: `blocker.created/updated/status_changed/deleted`
|
||||
|
||||
```text
|
||||
GET/POST /api/initiatives/{id}/blockers
|
||||
GET/PATCH/DELETE /api/blockers/{id}
|
||||
```
|
||||
|
||||
Capabilities: **`kairo.blocker.read`**, **`kairo.blocker.manage`**
|
||||
|
||||
## Frontend
|
||||
|
||||
Vorhaben-Detail: Sektion **Blocker**
|
||||
|
||||
---
|
||||
|
||||
# Teil C — AP0.8c: BacklogItem
|
||||
|
||||
## Migration (gleiche `007`)
|
||||
|
||||
```sql
|
||||
backlog_items (
|
||||
id UUID PK,
|
||||
tenant_id UUID NOT NULL,
|
||||
initiative_id UUID NOT NULL REFERENCES initiatives(id),
|
||||
title VARCHAR(255) NOT NULL,
|
||||
description TEXT DEFAULT '',
|
||||
status VARCHAR(32) NOT NULL, -- new, triaged, accepted, rejected, converted
|
||||
priority VARCHAR(16) DEFAULT 'normal',
|
||||
converted_action_id UUID NULL REFERENCES actions(id),
|
||||
created_at, updated_at
|
||||
)
|
||||
```
|
||||
|
||||
## Service + API
|
||||
|
||||
```text
|
||||
GET/POST /api/initiatives/{id}/backlog
|
||||
GET/PATCH/DELETE /api/backlog/{id}
|
||||
POST /api/backlog/{id}/convert-to-action
|
||||
```
|
||||
|
||||
Capabilities: **`kairo.backlog.read`**, **`kairo.backlog.manage`**
|
||||
|
||||
Convert: erzeugt Action, setzt `converted_action_id`, status=converted, Audit.
|
||||
|
||||
## Frontend
|
||||
|
||||
Vorhaben-Detail: Sektion **Backlog** + Button „In Maßnahme umwandeln“
|
||||
|
||||
---
|
||||
|
||||
# Teil D — AP0.8d: Milestone minimal
|
||||
|
||||
## Migration (gleiche `007`)
|
||||
|
||||
```sql
|
||||
milestones (
|
||||
id UUID PK,
|
||||
tenant_id UUID NOT NULL,
|
||||
initiative_id UUID NOT NULL REFERENCES initiatives(id),
|
||||
title VARCHAR(255) NOT NULL,
|
||||
goal_description TEXT DEFAULT '',
|
||||
status VARCHAR(32) NOT NULL, -- planned, active, at_risk, reached, moved, discarded
|
||||
target_date DATE NULL,
|
||||
created_at, updated_at
|
||||
)
|
||||
```
|
||||
|
||||
**Hinweis Zielarchitektur:** Langfristig `RoadmapItem(item_type=milestone)`. Diese Tabelle ist bewusste MVP-Brücke (AD-TA-08).
|
||||
|
||||
## Service + API
|
||||
|
||||
```text
|
||||
GET/POST /api/initiatives/{id}/milestones
|
||||
GET/PATCH/DELETE /api/milestones/{id}
|
||||
```
|
||||
|
||||
Capabilities: **`kairo.milestone.read`**, **`kairo.milestone.manage`**
|
||||
|
||||
Attention-Regel 6 nach Implementierung aktivieren.
|
||||
|
||||
## Frontend
|
||||
|
||||
Vorhaben-Detail: Sektion **Meilensteine** (kompakt, kein Gantt)
|
||||
|
||||
---
|
||||
|
||||
# Capabilities AP0.8
|
||||
|
||||
| Capability | Modul | Grants |
|
||||
|------------|-------|--------|
|
||||
| `kairo.attention.read` | attention | Member+ (wie workspace.read) |
|
||||
| `kairo.blocker.read` | blocker | Member+ |
|
||||
| `kairo.blocker.manage` | blocker | Member+ |
|
||||
| `kairo.backlog.read` | backlog | Member+ |
|
||||
| `kairo.backlog.manage` | backlog | Member+ |
|
||||
| `kairo.milestone.read` | milestone | Member+ |
|
||||
| `kairo.milestone.manage` | milestone | Member+ |
|
||||
|
||||
Registrierung: `backend/rights_registrations/operating_model_ops.py`
|
||||
Gesamt nach AP0.8: **25 Capabilities** (18 + 7)
|
||||
|
||||
---
|
||||
|
||||
# Architekturregeln (verbindlich)
|
||||
|
||||
1. **Tenant-first** — `tenant_id` aus TenantContext
|
||||
2. **Actor-first** — `reported_by_actor_id` optional; Assignments unverändert
|
||||
3. **Data Layer read / Services write** — Attention nur Data Layer
|
||||
4. **Router dünn** — keine Aggregations-SQL in Routern
|
||||
5. **Signals erklärbar** — reason_code, scope_type, data_source in DTOs
|
||||
6. **Cross-Tenant → 404**
|
||||
7. **Migration `007`** — kein ad-hoc DDL
|
||||
8. **Kein Prompt/KI/MCP/Workflow/SteeringContext/Method Registry**
|
||||
9. **Product Reset** — drei Ebenen RoadmapItem/BacklogItem/Action respektieren
|
||||
|
||||
---
|
||||
|
||||
# Nicht-Scope AP0.8
|
||||
|
||||
- Project / Program / SteeringContext / Method Registry / Hook Runtime
|
||||
- Evidence, Decision, Review, RecurringElement (→ AP0.9)
|
||||
- Action-Status `ready`, `review_required`
|
||||
- Due Dates auf Actions / Überfälligkeits-Engine
|
||||
- KI, MCP, Workflow Engine
|
||||
- Roadmap/RoadmapLane/RoadmapItem-Modell (→ Phase C)
|
||||
- Workspace-weite Backlog-Navigation
|
||||
- AP0.10 Validation Testvorhaben
|
||||
|
||||
---
|
||||
|
||||
# Tests (Mindestumfang)
|
||||
|
||||
## Backend — `backend/tests/test_ap08_operating_model.py`
|
||||
|
||||
| Bereich | Tests |
|
||||
|---------|-------|
|
||||
| Attention | Regeln 1, 1b, 2–6, Sortierung, Cross-Tenant |
|
||||
| NextAction | assign, convert_backlog, create_action, limit |
|
||||
| Blocker | CRUD, Tenant-Isolation, optional action_id |
|
||||
| Backlog | CRUD, convert-to-action |
|
||||
| Milestone | CRUD, at_risk in Attention |
|
||||
| Capabilities | Member grants, Gate 403, 25 sync |
|
||||
| Migration | `007` in test_migrations |
|
||||
|
||||
## Frontend
|
||||
|
||||
- Vitest: Registry enthält `kairo.attention`
|
||||
- Build grün
|
||||
|
||||
---
|
||||
|
||||
# Abnahmekriterien AP0.8
|
||||
|
||||
1. Migration `007`; Schema `007`
|
||||
2. Attention/NextAction regelbasiert (Regeln 1–6)
|
||||
3. Blocker, Backlog, Milestone CRUD tenant-sicher
|
||||
4. `data_layer/attention.py`; Router delegieren
|
||||
5. AttentionWidget im Workspace
|
||||
6. Vorhaben-Detail: Blocker, Backlog, Meilensteine
|
||||
7. Convert Backlog → Action
|
||||
8. Cross-Tenant-Tests grün
|
||||
9. 25 Capabilities registriert
|
||||
10. AP0.6b UX erhalten
|
||||
11. Kein Prompt/KI/MCP/Steering-Scope
|
||||
12. Abschlussbericht + README
|
||||
|
||||
---
|
||||
|
||||
# Abschlussbericht (nach Umsetzung)
|
||||
|
||||
Siehe Vorlage in v0.1 — Abschnitte 1–15.
|
||||
|
||||
---
|
||||
|
||||
# Freigegebene Punkte (ehemals offen)
|
||||
|
||||
| # | Entscheidung |
|
||||
|---|--------------|
|
||||
| 1 | Ein AP0.8, 4 interne Slices |
|
||||
| 2 | Eigene Capability `kairo.attention.read` |
|
||||
| 3 | Blocker/Action-Status optional synchron |
|
||||
| 4 | Convert Backlog→Action Pflicht |
|
||||
| 5 | Member manage wie Actions |
|
||||
| 6 | Initiative draft → AP0.8.1 |
|
||||
| 7 | Backlog nur Initiative-Detail |
|
||||
|
||||
---
|
||||
|
||||
*Freigegeben — Umsetzung gemäß Slice-Reihenfolge 8a→8d.*
|
||||
@@ -0,0 +1,149 @@
|
||||
# AP0.8 – Abschlussbericht Operating Model Extension I
|
||||
|
||||
**Status:** abgeschlossen (Implementierung) · QA mit DB ausstehend
|
||||
**Stand:** 2026-07-05
|
||||
**Version:** Backend/Frontend `0.8.0-ap0.8` · Schema `007`
|
||||
**Auftrag:** `Sprint0_AP0_8_Assignment_v0.2.md` (abgeglichen mit System Target State v0.1)
|
||||
|
||||
---
|
||||
|
||||
## 1. Scope und Einordnung
|
||||
|
||||
AP0.8 ist der erste fachliche Ausbau nach dem Product Reset. Er operationalisiert die Leitfrage über regelbasierte Attention/NextAction und erweitert das Domänenmodell um Blocker, BacklogItem und Milestone.
|
||||
|
||||
Einordnung in der Evolutionslinie (System Target State §24):
|
||||
|
||||
- Foundation ✓ (AP0.7)
|
||||
- **Signal / NextAction Engine — erste Instanz** ✓ (AP0.8a, Data Layer)
|
||||
- Operating Model Phase A ✓ (Blocker, Backlog, Milestone)
|
||||
- SteeringContext / Method Registry — bewusst **nicht** in AP0.8
|
||||
|
||||
---
|
||||
|
||||
## 2. Definition of Done — Prüfmatrix
|
||||
|
||||
| Kriterium | Ergebnis | Nachweis |
|
||||
|-----------|----------|----------|
|
||||
| Migration 007 | ✓ | `007_operating_model_extension_i.sql` |
|
||||
| Attention Regeln 1–6 | ✓ | `data_layer/attention.py` |
|
||||
| NextActionCandidates | ✓ | `get_next_action_candidates` |
|
||||
| Blocker CRUD | ✓ | `services/blockers.py`, Router |
|
||||
| Backlog CRUD + Convert | ✓ | `services/backlog.py` |
|
||||
| Milestone CRUD | ✓ | `services/milestones.py` |
|
||||
| AttentionWidget | ✓ | `widgets/AttentionWidget.jsx` |
|
||||
| Initiative-Detail Sektionen | ✓ | Blockers/Backlog/Milestones |
|
||||
| 25 Capabilities | ✓ | `operating_model_ops.py` |
|
||||
| Frontend Vitest + Build | ✓ | 9/9, build grün |
|
||||
| Backend pytest | ⏳ Remote | Gitea `test.yml` nach Push auf `develop` (Pi, Container) |
|
||||
| Kein KI/MCP/Steering | ✓ | Scope eingehalten |
|
||||
|
||||
---
|
||||
|
||||
## 3. Migration 007
|
||||
|
||||
Tabellen: `blockers`, `backlog_items`, `milestones` — alle mit `tenant_id`, FK RESTRICT/CASCADE wie AP0.5.
|
||||
|
||||
---
|
||||
|
||||
## 4. Attention / NextActionCandidate
|
||||
|
||||
- Datei: `backend/data_layer/attention.py`
|
||||
- API: `GET /api/workspace/attention`, `GET /api/workspace/next-actions`
|
||||
- Regeln: blocked_action, open_blocker, high_priority (personal), unassigned, initiative_without_next_action, stale_initiative, milestone_at_risk
|
||||
- DTOs mit `reason_code`, `scope_type`, `data_source` (Erklärbarkeit gemäß System Target State §17)
|
||||
|
||||
---
|
||||
|
||||
## 5. Blocker
|
||||
|
||||
- Service + nested/flat API
|
||||
- Audit: `blocker.created/updated/status_changed/deleted`
|
||||
- Optional: `set_action_blocked` bei Create
|
||||
|
||||
---
|
||||
|
||||
## 6. BacklogItem
|
||||
|
||||
- Getrennt von Actions (RoadmapItem/Backlog/Action-Unterscheidung)
|
||||
- Convert-to-action mit Audit `backlog.converted_to_action`
|
||||
|
||||
---
|
||||
|
||||
## 7. Milestone minimal
|
||||
|
||||
- MVP-Brücke (langfristig RoadmapItem — AD-TA-08)
|
||||
- Attention-Regel 6 für `at_risk`
|
||||
|
||||
---
|
||||
|
||||
## 8. Data Layer Erweiterung
|
||||
|
||||
Neu: `attention.py` — Read-only, tenant-scoped, keine Router-SQL.
|
||||
|
||||
---
|
||||
|
||||
## 9. API-Endpunkte
|
||||
|
||||
Siehe README AP0.8-Tabelle.
|
||||
|
||||
---
|
||||
|
||||
## 10. Capabilities
|
||||
|
||||
7 neue: `kairo.attention.read`, `kairo.blocker.*`, `kairo.backlog.*`, `kairo.milestone.*`
|
||||
Gesamt: 25
|
||||
|
||||
---
|
||||
|
||||
## 11. Frontend-Integration
|
||||
|
||||
- `AttentionWidget` (order 8, capability `kairo.attention.read`)
|
||||
- `BlockersSection`, `BacklogSection`, `MilestonesSection` im Vorhaben-Detail
|
||||
- Status-Badges erweitert für blocker/backlog/milestone
|
||||
|
||||
---
|
||||
|
||||
## 12. Tests und Verifikation
|
||||
|
||||
| Test | Status |
|
||||
|------|--------|
|
||||
| `test_ap08_operating_model.py` | 14 Tests — Lauf auf Pi via `test.yml` |
|
||||
| `test_rights_registry.py` | aktualisiert (25 Capabilities) |
|
||||
| `test_migrations.py` | Migration 007 ergänzt |
|
||||
| `conftest.py` | Session-Guard: skip lokal ohne DB, kein Migrations-Timeout |
|
||||
| Vitest | 9/9 grün (lokal) |
|
||||
| `vite build` | grün (lokal) |
|
||||
|
||||
**Verifikation:** Push auf `develop` → `deploy-dev.yml` + `test.yml` (pytest im Backend-Container gegen PostgreSQL im Compose-Stack).
|
||||
|
||||
---
|
||||
|
||||
## 13. Abweichungen von diesem Auftrag
|
||||
|
||||
| Punkt | Abweichung |
|
||||
|-------|------------|
|
||||
| Auftrag v0.1 | Ersetzt durch v0.2 nach Abgleich System Target State |
|
||||
| Attention DTO | Zusätzlich `scope_type`, `data_source`, Regel 1b open_blocker |
|
||||
| NextAction DTO | Explizit dokumentiert und implementiert |
|
||||
| Milestone API target_date | Router akzeptiert ISO-Date-String |
|
||||
|
||||
Keine fachlichen Scope-Abweichungen.
|
||||
|
||||
---
|
||||
|
||||
## 14. Offene Punkte für AP0.9
|
||||
|
||||
- Evidence, Decision, Review, RecurringElement
|
||||
- Action-Status `ready`, `review_required`
|
||||
- Due Dates auf Actions + Attention Regeln 7–9
|
||||
- Attention-Refactor in `backend/steering/signals/` (wenn Steering Core startet)
|
||||
|
||||
---
|
||||
|
||||
## 15. Empfehlung für AP0.10 / Validation
|
||||
|
||||
Nach AP0.9 ein reales Testvorhaben durchspielen (Backlog → Action → Blocker → Meilenstein → Attention). Milestone→RoadmapItem-Konsolidierung als ADP vor Phase C klären.
|
||||
|
||||
---
|
||||
|
||||
*Implementierung abgeschlossen — Backend-pytest und Smoke auf dem Pi nach Push auf `develop`.*
|
||||
Reference in new issue
Block a user