mbse-system-modelinglisted

# MBSE System Modeling Use this skill when the goal is to understand the current system design through multiple views and abstraction levels, not to propose a future architecture first. ## When to Use - System understanding is harder than implementation. - Responsibilities, scopes, or subsystem boundaries feel ambiguous. - You need diagrams and structured views of the current design. - You want to reveal overlaps, duplicated control, or mixed abstraction levels. - A refactor or RFC needs an evidence-backed model of the system as it exists. ## Core Principle Model the system **as-is**, not as intended. Start broad, then decompose only the hotspots. Keep filenames and section headings stable so later runs update existing documents instead of inventing new ones. ## Output Layout All generated or updated docs live under: `docs/system-models/` Read [references/layout.md](references/layout.md) before writing. It defines the exact tree, file names, and required sections. If the scaffold does not exist yet, initialize it: ```bash python3 "${CLAUDE_SKILL_DIR}/scripts/init_mbse_docs.py" --root "$(git rev-parse --show-toplevel)" ``` To scaffold specific hotspot subsystems: ```bash python3 "${CLAUDE_SKILL_DIR}/scripts/init_mbse_docs.py" \ --root "$(git rev-parse --show-toplevel)" \ --subsystem supervisor-agent \ --subsystem agent-adapter \ --subsystem attention-router ``` ## Modeling Standards Read [references/standards.md](references/standards.md) before drafting