菜单
Recommend the next one to three concrete actions with the highest leverage for the latest actionable result or current Binding. Use only when the user invokes next or asks what to do next; never redirect or act silently.
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
Supply the shared Think It Through deck rules and mental model whenever the user invokes deck, plays or combines any Think It Through card, invokes help, or asks to use the deck. Stay silent while a card or help utility produces the response; when invoked alone, initialize the deck on the available conversation.
Explore and deepen the thought currently being expressed as an active, neutral thinking partner without forcing a conclusion or changing effects. Use only when the user invokes discuss or asks to reason through something together; never redirect silently.
Separate and clarify one or more fragmented, implicit, or hard-to-word thoughts, expose only supported relationships, then respond or pass the result forward. Use only when the user invokes distill or asks for clarification before answering; never activate it silently.
Explain the current HACP Working Object at the user's requested or apparent level with the minimum detail needed for understanding while preserving its claims and caveats. Use when the user invokes explain, asks for an explanation, or passes a result from another deck into EXPLAIN; never add an explanation silently.
Push the current thought or HACP Working Object one grounded creative leap beyond its present edge without broad brainstorming, choosing a direction, or inventing support. Use only when the user invokes further or explicitly asks to push the current idea further.
Stress-test a proposal, assumption, decision, design, or plan through a relentless multi-turn grill that follows its decision tree one question and recommendation at a time. Use only when the user invokes grill or asks for a demanding challenge; never begin silently.
| name | next |
| description | Recommend the next one to three concrete actions with the highest leverage for the latest actionable result or current Binding. Use only when the user invokes next or asks what to do next; never redirect or act silently. |
Use when: The user has enough clarity to continue but needs the highest-leverage next step. Default binding: The latest actionable result, otherwise the current Binding. Accepts: A compatible HACP Working Object or the declared default material. Effect: Recover the current stage and dependencies, identify the bottleneck, then rank concrete actions by leverage. Result: One to three actions with expected outcomes. Duration: One agent turn. Limits: Distinguish conversational and external actions. Do not expand into a full plan or execute anything.
flowchart LR
A["Actionable result or Binding"] --> B["Recover stage and dependencies"]
B --> C["Find highest-leverage bottleneck"]
C --> D["Rank possible actions"]
D --> E["Select one to three"]
E --> F["Name action and expected outcome"]
F --> G["User chooses whether to act"]
Prefer a reversible learning step when uncertainty is high.
Begin the combo trace with > 🎯 **<binding>** → ⚡ **NEXT**, followed by one to three Next actions ordered by leverage.
Add an output with → and presentation cards with +; show the trace once for the complete combo.