cs-design-sample1

🤝 Partnership — /cs-partnership with [skills/agents] [goal]. v5.5: AGENT type (executor, debugger, architect) directly calls external plugin agents via Task(subagent_type). SKILL/PROTOCOL types use Skill() or direct execution.

🚢 DevOps — Pre-PR validation gate. 4-agent team checks spec compliance, coverage, commit-message quality before merge.

🗣️ Language Coach — Convert session Q&A into natural American English conversations for language learning.

⚡ Team Lead — Two-phase orchestrator: plan with Opus → execute with Sonnet agents in parallel.

🎯 Goal — /goal with Goal Gate built-in. Clarifies goal first, then auto-dispatches to the best CS-series domain via cs-ceo.

🔍 Code Reviewer — 5-agent parallel review: Architecture, Quality, Security, Performance, Maintainability.

📚 Knowledge Keeper — Meta-router across CS-test / CS-plan / CS-codebase-review / cs-design with versioned learnings. Includes /cs-end session-wrap.

🧭 CEO — Estimates effort and autonomously dispatches across CS-test/CS-plan/CS-codebase-review/cs-design/cs-smart-run. v5.5 adds Dynamic Resolve v2: partner type auto-detection (AGENT/SKILL/PROTOCOL) + external agent direct call support.

🏁 Session Closer — 4-Agent analysis + TIL save + error note capture (Phase 2.2) + plugin version-up + GitHub push. Self-contained, no external vault. (/cs-end)

장기 핵심 메모리 — 세션 간 전략적 지혜 축적 ('30년 근속' 맥락 제공)