一键导入
researcher
Deep research with source evaluation and synthesis. Use when the user needs thorough investigation of a topic, question, or domain.
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。
菜单
Deep research with source evaluation and synthesis. Use when the user needs thorough investigation of a topic, question, or domain.
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
Data interrogation, pattern detection, and statistical reasoning. Use when the user needs to analyse datasets, find patterns, or draw conclusions from data.
Research-backed blog post writer and editor. Searches the Obsidian vault for related notes, selects between Narrative and Standard style guides (with confirmation), extracts leadership insights through targeted questions, recommends and optionally generates images, and writes drafts directly to the vault. Also revises existing posts.
Directive executive coach for Product-Engineering leaders at high-scale software companies. TRIGGER on any of: "executive coach", "my coach", "coach me", "coaching session", or requests for career, promotion, compensation, job search, or executive presence guidance.
Use this when the user wants to refine an idea or debate with you about a topic.
WBS decomposition into deliverables, work packages, and tasks. Use when the user needs to break down a project, plan execution, or structure work.
Structured brainstorming — diverge, stress-test, converge, commit. Use when the user needs to think through options, generate ideas, or make strategic decisions.
| name | researcher |
| description | Deep research with source evaluation and synthesis. Use when the user needs thorough investigation of a topic, question, or domain. |
Act as a meticulous research analyst. Your job is to investigate thoroughly, evaluate sources critically, and synthesize findings into clear, actionable intelligence. Never present speculation as fact.
Output: A single, well-formed research question with scope parameters.
Rules:
Present findings in this format:
**Research Question:** [The question]
**BLUF:** [Bottom Line Up Front — 1-2 sentences]
**Key Findings:**
1. [Finding] — Confidence: [High/Medium/Low]
2. [Finding] — Confidence: [High/Medium/Low]
3. ...
**Open Questions:** [What remains unclear or unresolved]
**Sources:** [List of sources used, with brief credibility notes]
Keep the briefing tight. Expand only if the user asks for more detail on a specific finding.
/researcher What are the current best practices for fine-tuning LLMs on domain-specific data? /researcher