菜单
Keep Ralph operational files internally consistent with single-source decision logging. Use when updating STATE, decisions, learnings, and task trackers.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Generate EventModel artifacts from ADDR phase deliverables (Align/Define/Design/Refine). Use when the user asks to create or refresh EventModel slices from ADDR artifacts, API profiles, or OpenAPI files, and when domain commands/events/read models must be derived from API operations.
Generate top-down schematic railway track diagrams as draw.io (.drawio) XML files. LEGO/BRIO-style snap-together pieces using turtle graphics — straight, bend, wide bend, turnout, merger. Use when the user asks to create railway track diagrams, causal flow topology diagrams with track metaphors, or draw.io files with railway-style layouts.
Analyze legacy system architectures and produce Event Modeling slice models in JSON format. Transforms system descriptions into business-focused Event Models with commands, events, read models, screens, and automations. Use when the user asks to model a legacy system, create event models, generate Event Sourcing slices, or convert system behavior into Event Modeling JSON.
Generate Big Picture Event Storming diagrams as draw.io (.drawio) XML files with proper styling. Handles element shapes, colors, pivotal markers, boundary ellipses, and programmatic XML generation. Use when the user asks to create, generate, or update a draw.io event storming diagram, or when working with .drawio files for event storming.
Domain knowledge for Big Picture Event Storming following DDD conventions. Covers element types, color coding, pivotal event placement, layout rules, and bounded context identification. Use when the user asks about event storming, domain events, bounded contexts, DDD modeling, Big Picture EventStorming, pivotal events, or event storming workshop conventions.
| name | ralph-loop-governance |
| description | Keep Ralph operational files internally consistent with single-source decision logging. Use when updating STATE, decisions, learnings, and task trackers. |
Use this skill when updating any files under .ralph/.
Decisions have exactly one source of truth:
.ralph/agent/decisions.md.ralph/STATE.md.ralph/STATE.md; mark all others as baseline/reference.ralph/STATE.md: handoff status only (phase, progress, blockers, next steps, pointers).ralph/agent/decisions.md: all formal decisions with rationale.ralph/agent/learnings.md: LRN-XXX learnings and actionable prevention notes.ralph/tasks/*.md: task checklists and execution progressBefore commit:
STATE.md references decisions.md but does not repeat decision contentdecisions.mdlearnings.md