AgentSkillsCN

event-modeling-brownfield

从现有代码库中逆向工程事件模型，以梳理并记录隐性事件流转。

SKILL.md
--- frontmatter
name: event-modeling-brownfield
description: Reverse engineer Event Model from existing codebase to document implicit event flows.
argument-hint: <codebase-path-or-description>
disable-model-invocation: true

Reverse engineer the Event Model that "should have existed" for a brownfield codebase.

This command loads all relevant preferences for: Event Modeling (Qlerify 7-step methodology), functional domain modeling (DDD aggregates, Decider pattern), bounded context design (context mapping, ACL), event sourcing (CQRS, event replay, state reconstruction), strategic domain analysis (Core/Supporting/Generic classification), EventCatalog transformation (schema documentation), railway-oriented programming (Result types), and algebraic laws (property-based testing).

Workflow

Phase 1: Codebase Analysis

If $ARGUMENTS provided, use that path or description. Otherwise, ask the user to specify the codebase location or describe the system.

Analyze the codebase to identify:

  • •Existing domain entities and their state transitions
  • •API endpoints or UI actions that trigger state changes
  • •Database schemas and their relationships
  • •Error handling patterns and validation rules
  • •Integration points with external systems

Phase 2: Event Extraction (Reverse Step 1)

Extract implicit domain events from:

  • •Database writes and state mutations
  • •Log statements indicating business actions
  • •Method names suggesting state changes (create, update, delete, approve, reject)
  • •Audit trails or history tables

Name events in past tense: "OrderPlaced", "PaymentReceived", "UserRegistered"

Phase 3: Command Identification (Reverse Steps 3-4)

Identify commands from:

  • •API endpoint handlers
  • •UI form submissions
  • •Message queue consumers
  • •Scheduled job triggers

Map commands to the events they produce. Identify validation rules as GWT preconditions.

Phase 4: Read Model Discovery (Reverse Step 5)

Identify read models from:

  • •Database queries preceding writes
  • •API responses informing UI state
  • •Cached or denormalized data structures

Phase 5: Bounded Context Analysis (Reverse Step 6)

Apply strategic domain analysis to classify:

  • •Core domains (competitive advantage, high investment)
  • •Supporting domains (necessary but not differentiating)
  • •Generic domains (commodity, buy or use standard solutions)

Identify bounded context boundaries from:

  • •Module/package structure
  • •Database schema separation
  • •Team ownership patterns
  • •Integration points requiring translation

Phase 6: GWT Scenario Extraction (Reverse Step 7)

Extract Given-When-Then scenarios from:

  • •Existing test cases
  • •Validation error messages
  • •Business rule documentation
  • •Edge case handling code

Phase 7: Artifact Generation

Produce:

  • •Event Model documentation suitable for EventCatalog
  • •Bounded context map
  • •Aggregate identification with Decider pattern mapping
  • •Gap analysis: what's missing vs ideal Event Model

User Input

$ARGUMENTS