메뉴
Use when the user asks about "thinking modes", "effort levels", "how hard should Claude think", or wants to pick between low/medium/high/xhigh/max for a task — explains the cost/quality tradeoff per level and recommends one based on the work shape.
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
Use when the user describes a multi-step feature or refactor and you need to decide whether to handle it solo, dispatch a parallel `/fan-out`, or run a `/worktree-team` planner→generator→reviewer pipeline. Outputs a routing recommendation with the reasoning behind it.
Use whenever the user asks to apply the same operation to many independent files or components — bulk migration, batch refactor across controllers, parallel exploration of unrelated subsystems. Dispatches one subagent per file with a shared prompt template, then collects the results.
Use to toggle the safe-mode flag at .claude/safe-mode. The flag is written automatically by consecutive-failure-guard once tool failures hit FORGE_SAFE_MODE_THRESHOLD (default 5), and block-destructive.sh denies every Bash/Write/Edit while it exists. /safe-mode off clears the flag, logs the exit, and prompts /postmortem; /safe-mode on enters manually; /safe-mode status reports current state.
Run the 5-stage Token Transformation Pipeline (TRAE §5.2.2) over the current session state — Collection, Ranking, Compression, Budgeting, Assembly. Emits a concrete next-action recommendation (/compact, /lean-md, or /progress-log + fresh session) instead of a generic warning.
Audit or scaffold a CLAUDE.md against the Karpathy 4-section structure (Think Before Coding · Simplicity First · Surgical Changes · Goal-Driven Execution). Companion to /lean-md — structure audit, not size trim.
Print every policy enforcement point declared in plugins/diagnostics/registry/policies.json — id, verdict, plugin, hook event, severity, bypass — grouped by verdict. Single discoverable index of what the harness blocks, anchors, nudges, or logs at runtime.
| name | ultrathink |
| description | Use when the user asks about "thinking modes", "effort levels", "how hard should Claude think", or wants to pick between low/medium/high/xhigh/max for a task — explains the cost/quality tradeoff per level and recommends one based on the work shape. |
| when_to_use | Reach for this when picking an `effort:` value for a new skill or `/effort` for the current turn, or when teaching the reasoning modes. Do NOT use for changing the effort level — use the `/effort` command instead; ultrathink is the explanation, not the switch. |
| disable-model-invocation | true |
| model | haiku |
| allowed-tools | [] |
| logical | explanation of effort levels emitted with cost/quality tradeoff per level |
Control how hard Claude thinks with /effort or the CLAUDE_CODE_EFFORT_LEVEL env var:
| Level | When to use | Token cost |
|---|---|---|
| low | Simple lookups, quick questions, high-volume tasks | Minimal |
| medium | Standard development, balanced speed/quality | Moderate |
| high | Complex reasoning, nuanced analysis, difficult bugs (default) | Higher |
| xhigh | Hardest tasks needing deeper reasoning; persists in settings | High |
| max | Deepest reasoning, session-only (set per-turn or via env) | Highest |
Include "ultrathink" anywhere in your prompt to request deeper reasoning on that turn. Claude Code recognizes the keyword and adds an in-context instruction to think harder — your session effort level is unchanged. No need to change global settings.
Example: "ultrathink: design the auth system for this app considering OAuth, JWT, and session-based approaches"
ultracode is a separate keyword. ultrathink adds an in-context "think harder" instruction for one turn; ultracode (or /effort ultracode) tells Claude Code to plan a native dynamic workflow — a background script that orchestrates many subagents with verification built in — for the task. It pairs xhigh effort with automatic workflow orchestration, lasts the session, and uses more tokens per task. Reach for ultrathink to reason harder in place; reach for ultracode to fan a large task out to a verified multi-agent workflow. It appears in the /effort menu only on Claude Code builds that support dynamic workflows — if you don't see it there, the feature isn't available.
Claude dynamically decides how much to think based on query complexity. Higher effort + harder query = more thinking. Easy queries get direct responses even at high effort — no token waste.
Use the Monitor tool to stream events from background scripts in real time. Useful for watching long-running subagent output without polling.
/effort to change level