메뉴
Use when stuck on hard problems, need deeper reasoning, standard approaches fail, or facing architecture decisions. Implements 5-tier escalation from standard reasoning to external expert models.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Run a multi-agent code review on a GitHub pull request, with parallel security / performance / architecture / style passes, inline comments, and quality-gate enforcement. Use when asked to "github code review", "review github pr", "run code review on PR", "swarm review", "automated PR review", or "post inline review comments".
Manage GitHub issues, project boards, sprints, and milestones with swarm-coordinated automation — issue triage, decomposition into subtasks, board sync, sprint planning, KPI tracking. Use when asked to "github project management", "manage github projects", "github issue automation", "swarm project board", "automate sprint", "triage issues", or "set up project board".
Extract clean article content from a URL (blog post, news article, tutorial) and save it as readable text with ads, navigation, and other clutter removed. Use when asked to "extract article", "scrape text from URL", "download blog post", "parse HTML article", "save article as text", or "clean article content from a webpage".
Dispatch one focused agent per independent problem domain so investigations run concurrently. Use when facing 2+ independent tasks with no shared state — for example "spawn agents in parallel", "parallel debugging", "fan out tasks", "investigate multiple failures at once", or "distribute independent work across agents".
Create HTML dashboards with KPI cards, bar/pie/line charts, progress indicators, and data tables. Use when asked to "create dashboard", "build KPI dashboard", "generate metrics visualization", "make analytics dashboard", "show performance dashboard", or "render data visualization as HTML".
Guide for choosing the right AI coding tool for the task. Use when user asks "which AI should I use", "should I use Claude or Cursor", "what's the best tool for X", or when helping decide between Claude Code, Cursor, Windsurf, VSCode + Copilot, Gemini, Grok, Codex, Qwen, or this Claude interface.
| name | escalation-reasoning |
| description | Use when stuck on hard problems, need deeper reasoning, standard approaches fail, or facing architecture decisions. Implements 5-tier escalation from standard reasoning to external expert models. |
| Tier | Method | Reasoning Depth | Best For |
|---|---|---|---|
| 1 | Standard | Default | Routine tasks, simple fixes |
| 2 | /think | 4K tokens | Debugging, code reviews |
| 3 | /megathink | 10K tokens | Architecture decisions |
| 4 | /ultrathink | 32K tokens | Hard problems, novel approaches |
| 5 | External | Expert models | True blockers, breakthroughs |
/think)/megathink)/ultrathink)| Model | Strength | Cost | Best For |
|---|---|---|---|
| o3 | Reasoning | $$$$ | Hard math/logic |
| Gemini 2.5 Pro | Context | $$$ | Large codebases |
| DeepSeek-R1 | Value | $ | Cost-effective reasoning |
| Codex | Code | $$ | Pure implementation |
DeepSeek is 97% cheaper than equivalent Claude reasoning.
Before escalating, ask:
# Tier 2
/think: [your question]
# Tier 3
/megathink: [your question]
# Tier 4
/ultrathink: [your question]
# Tier 5 - Use consult-llm tool
Consult o3 about: [your question]