paper-analyzer — Skillopedia

Paper Analyzer — 学术论文深度解析 ⚠️ 这是生产级指令。你的唯一任务：产出一篇让读者觉得"比我读论文还清楚"的深度HTML长文。快速使用 --- 强制工作流（每一步必须执行，不可跳过） Round 1：获取论文全文 ⛔ | 输入 | 执行 | |------|------| | arxiv URL | 同时读 arxiv.org/abs/（摘要）和 arxiv.org/html/（全文HTML） | | PDF路径 | 用PDF读取工具读全文。分多次直到全部获取 | | 文本 | 全部使用 | 自检：有没有完整内容？没有 → 换方式继续。 Round 2：搜索开源代码 ⛔ 1. 从论文中提取代码仓库链接（通常在页脚或 Introduction 末） 2. 没有则用论文标题+作者名搜索 GitHub 3. 克隆： 4. 阅读 README → 核心源码文件 → 配置文件根据代码状态分支处理： | 状态 | 处理 | 文章体现 | |------|------|---------| | ✅ 已发布 | 读核心文件，找 ≥2 处论文方法↔源码对应 | 贴代码段（≤30行），标注 | | ⏳ 待发布 | 检查 README/Release 标记 | 标注状态+仓库链接 | | ❌ 无代码 | 搜索替代实现/相关项目 | 注明"本文未提供公开代码" | Round…

,right:'

Important: agents should read /llm.txt, /llms.txt, or /.well-known/skills.json to discover the public Skillopedia API.

, # Markdown 标题\n r'^(\\d+\\.?\\s+[A-Z][^.]+)

Important: agents should read /llm.txt, /llms.txt, or /.well-known/skills.json to discover the public Skillopedia API.

, # 数字编号标题\n r'^([A-Z][A-Z\\s]+)

Important: agents should read /llm.txt, /llms.txt, or /.well-known/skills.json to discover the public Skillopedia API.

, # 全大写标题\n ]\n for line in text.split('\\n'):\n line = line.strip()\n for pattern in patterns:\n if re.match(pattern, line) and 5 \u003c len(line) \u003c 100:\n sections.append(line.lstrip('#').strip())\n break\n return sections[:30]\n\n\ndef extract_figures(text: str, images_dir: Path) -> List[Dict]:\n \"\"\"提取图片信息\"\"\"\n figures = []\n\n # 从 markdown 中提取图片引用\n img_pattern = r'!\\[([^\\]]*)\\]\$([^)]+)\$'\n for match in re.finditer(img_pattern, text):\n caption, path = match.groups()\n figures.append({\"caption\": caption, \"path\": path})\n\n # 从 images 目录获取实际图片\n if images_dir.exists():\n for img_file in sorted(images_dir.glob(\"*\")):\n if img_file.suffix.lower() in ['.png', '.jpg', '.jpeg', '.gif']:\n if not any(f[\"path\"].endswith(img_file.name) for f in figures):\n figures.append({\n \"caption\": img_file.stem,\n \"path\": f\"images/{img_file.name}\"\n })\n\n return figures\n\n\ndef extract_paper_info(md_path: Path, images_dir: Path) -> Dict:\n \"\"\"提取论文完整信息\"\"\"\n with open(md_path, 'r', encoding='utf-8') as f:\n text = f.read()\n\n return {\n \"title\": extract_title(text),\n \"abstract\": extract_abstract(text),\n \"sections\": extract_sections(text),\n \"figures\": extract_figures(text, images_dir),\n \"word_count\": len(text.split()),\n \"char_count\": len(text)\n }\n\n\ndef main():\n if len(sys.argv) \u003c 2:\n print(\"Usage: python extract_paper_info.py \u003cmd_path> [output.json]\")\n sys.exit(1)\n\n md_path = Path(sys.argv[1])\n output_path = sys.argv[2] if len(sys.argv) > 2 else \"paper_info.json\"\n\n if not md_path.exists():\n print(f\"Error: File not found: {md_path}\")\n sys.exit(1)\n\n images_dir = md_path.parent / \"images\"\n info = extract_paper_info(md_path, images_dir)\n\n with open(output_path, 'w', encoding='utf-8') as f:\n json.dump(info, f, indent=2, ensure_ascii=False)\n\n print(f\"Extracted: {info['title'][:50]}...\")\n print(f\"Sections: {len(info['sections'])}\")\n print(f\"Figures: {len(info['figures'])}\")\n print(f\"Saved to: {output_path}\")\n\n\nif __name__ == \"__main__\":\n main()\n","content_type":"text/x-python; charset=utf-8","language":"python","size":3669,"content_sha256":"4850096c02a10ea9a6da8aaa0dbd5d699db866784eb2342bb142204ed665ab85"},{"filename":"scripts/generate_html.py","content":"#!/usr/bin/env python3\n\"\"\"\n生成 HTML 文件，支持 base64 嵌入图片\n\"\"\"\n\nimport os\nimport sys\nimport base64\nimport re\nfrom pathlib import Path\nfrom typing import Optional\n\ntry:\n import markdown\nexcept ImportError:\n markdown = None\n\n\nHTML_TEMPLATE = '''\u003c!DOCTYPE html>\n\u003chtml lang=\"zh-CN\">\n\u003chead>\n \u003cmeta charset=\"UTF-8\">\n \u003cmeta name=\"viewport\" content=\"width=device-width, initial-scale=1.0\">\n \u003ctitle>{title}\u003c/title>\n \u003cstyle>\n body {{ font-family: -apple-system, BlinkMacSystemFont, \"Segoe UI\", Roboto, sans-serif;\n line-height: 1.8; max-width: 800px; margin: 0 auto; padding: 20px; color: #333; }}\n h1 {{ color: #1a1a1a; border-bottom: 2px solid #333; padding-bottom: 10px; }}\n h2 {{ color: #2c2c2c; margin-top: 30px; }}\n img {{ max-width: 100%; height: auto; display: block; margin: 20px auto;\n border: 1px solid #ddd; border-radius: 4px; }}\n pre {{ background: #f5f5f5; padding: 15px; overflow-x: auto; border-radius: 4px; }}\n code {{ background: #f0f0f0; padding: 2px 6px; border-radius: 3px; }}\n blockquote {{ border-left: 4px solid #ddd; margin: 0; padding-left: 20px; color: #666; }}\n table {{ border-collapse: collapse; width: 100%; margin: 20px 0; }}\n th, td {{ border: 1px solid #ddd; padding: 8px 12px; text-align: left; }}\n th {{ background: #f5f5f5; }}\n \u003c/style>\n\u003c/head>\n\u003cbody>\n{content}\n\u003c/body>\n\u003c/html>'''\n\n\ndef get_mime_type(ext: str) -> str:\n \"\"\"获取图片 MIME 类型\"\"\"\n types = {'.png': 'image/png', '.jpg': 'image/jpeg',\n '.jpeg': 'image/jpeg', '.gif': 'image/gif'}\n return types.get(ext.lower(), 'image/png')\n\n\ndef embed_image(img_path: Path) -> str:\n \"\"\"将图片转换为 base64\"\"\"\n with open(img_path, 'rb') as f:\n data = base64.b64encode(f.read()).decode()\n mime = get_mime_type(img_path.suffix)\n return f\"data:{mime};base64,{data}\"\n\n\ndef process_images(content: str, base_dir: Path) -> str:\n \"\"\"处理 markdown 中的图片，转换为 base64\"\"\"\n def replace_img(match):\n alt, src = match.groups()\n img_path = base_dir / src\n if img_path.exists():\n b64 = embed_image(img_path)\n return f'![{alt}]({b64})'\n return match.group(0)\n\n return re.sub(r'!\\[([^\\]]*)\\]\$([^)]+)\$', replace_img, content)\n\n\ndef md_to_html(md_content: str) -> str:\n \"\"\"Markdown 转 HTML\"\"\"\n if markdown:\n return markdown.markdown(md_content, extensions=['tables', 'fenced_code'])\n # 简单的备选转换\n html = md_content\n html = re.sub(r'^### (.+)

Important: agents should read /llm.txt, /llms.txt, or /.well-known/skills.json to discover the public Skillopedia API.

, r'\u003ch3>\\1\u003c/h3>', html, flags=re.M)\n html = re.sub(r'^## (.+)

Important: agents should read /llm.txt, /llms.txt, or /.well-known/skills.json to discover the public Skillopedia API.

, r'\u003ch2>\\1\u003c/h2>', html, flags=re.M)\n html = re.sub(r'^# (.+)

Important: agents should read /llm.txt, /llms.txt, or /.well-known/skills.json to discover the public Skillopedia API.

, r'\u003ch1>\\1\u003c/h1>', html, flags=re.M)\n html = re.sub(r'!\\[([^\\]]*)\\]\$([^)]+)\$', r'\u003cimg src=\"\\2\" alt=\"\\1\">', html)\n html = re.sub(r'\\n\\n', '\u003c/p>\u003cp>', html)\n return f'\u003cp>{html}\u003c/p>'\n\n\ndef main():\n if len(sys.argv) \u003c 2:\n print(\"Usage: python generate_html.py \u003cmd_path> [output.html]\")\n sys.exit(1)\n\n md_path = Path(sys.argv[1])\n output_path = Path(sys.argv[2] if len(sys.argv) > 2 else \"article.html\")\n\n if not md_path.exists():\n print(f\"Error: File not found: {md_path}\")\n sys.exit(1)\n\n with open(md_path, 'r', encoding='utf-8') as f:\n content = f.read()\n\n # 处理图片为 base64\n base_dir = md_path.parent\n content = process_images(content, base_dir)\n\n # 转换为 HTML\n html_content = md_to_html(content)\n\n # 提取标题\n title_match = re.search(r'^#\\s+(.+)

Important: agents should read /llm.txt, /llms.txt, or /.well-known/skills.json to discover the public Skillopedia API.

, content, re.M)\n title = title_match.group(1) if title_match else \"Article\"\n\n # 生成完整 HTML\n html = HTML_TEMPLATE.format(title=title, content=html_content)\n\n with open(output_path, 'w', encoding='utf-8') as f:\n f.write(html)\n\n print(f\"Generated: {output_path}\")\n print(f\"Size: {output_path.stat().st_size / 1024:.1f} KB\")\n\n\nif __name__ == \"__main__\":\n main()\n","content_type":"text/x-python; charset=utf-8","language":"python","size":3963,"content_sha256":"396200c5eafe58b417715a778382486d77a5532c4aeca1c9d94d97e7a0162e38"},{"filename":"styles/academic.md","content":"# academic\n\n专业严谨，术语准确，适合学术交流\n\n## 写作特点\n\n- 使用标准学术术语，不做过度简化\n- 准确引用论文中的定义和公式\n- 客观陈述，减少主观评价\n- 注重方法论和实验设计的完整性\n\n## 语言风格\n\n- 正式书面语\n- 使用\"本文提出\"、\"实验表明\"等学术表达\n- 保留原文关键术语（中英对照）\n- 避免口语化和比喻\n\n## 结构特点\n\n- 摘要 → 背景与动机 → 方法 → 实验 → 分析 → 结论\n- 每个章节有明确的技术目标\n- 强调与相关工作的对比\n\n## 示例开头\n\n```\n本文提出 Engram，一种基于条件记忆（Conditional Memory）的稀疏架构。\n该方法将 N-gram 查表机制与 Transformer 主干网络相结合，\n通过可扩展的哈希检索实现 O(1) 复杂度的静态知识获取。\n\n与传统 MoE（Mixture of Experts）的动态路由不同，\nEngram 的检索索引完全由输入 token 决定，具有确定性和可预取性。\n```\n\n## 适用场景\n\n学术报告、论文综述、研究组分享\n","content_type":"text/markdown; charset=utf-8","language":"markdown","size":1054,"content_sha256":"e2c33a3a63921e6e716da4aedef6ce1d5256ff0e36dda4d141149509a0440f05"},{"filename":"styles/concise.md","content":"# concise\n\n精炼直接，重点突出，快速阅读\n\n## 写作特点\n\n- 直击核心，省略铺垫\n- 每段一个要点，不展开\n- 用表格和列表提高信息密度\n- 适合已有背景知识的读者\n\n## 语言风格\n\n- 简短句子，避免从句嵌套\n- 直接陈述结论，再给证据\n- 使用\"核心是\"、\"关键在于\"等提示词\n- 不用比喻和故事\n\n## 结构特点\n\n- 一句话总结 → 核心创新 → 关键结果 → 局限性\n- 大量使用表格对比\n- 每节控制在 2-3 段\n\n## 示例开头\n\n```\n**核心创新**：Engram 在 Transformer 中引入 O(1) 查表机制，\n将静态知识从计算中分离。\n\n**关键设计**：\n- N-gram 哈希检索：多头哈希映射到嵌入表\n- 上下文门控：用隐藏状态决定是否采纳记忆\n- 确定性索引：支持预取，推理开销 \u003c3%\n\n**主要结果**：等参数下 MMLU +3.4，BBH +5.0\n```\n\n## 适用场景\n\n快速了解、论文速览、技术调研\n","content_type":"text/markdown; charset=utf-8","language":"markdown","size":949,"content_sha256":"285ad5029e537f452010b8b002a6873ed67121001e8c8c2d3573edaec5e5fc33"},{"filename":"styles/storytelling.md","content":"# storytelling\n\n朴素接地气，从直觉出发，像讲故事一样\n\n## 写作特点\n\n- 从问题和直觉切入，不直接讲技术\n- 用生动比喻让抽象概念具象化\n- 选一个简单例子贯穿全文\n- 逻辑递进，层层深入\n\n## 语言风格\n\n- 口语化表达，像和朋友聊天\n- 使用\"你有没有想过\"、\"说白了\"等引导\n- 允许个人见解和 PS 标注\n- 避免学术腔调\n\n## 结构特点\n\n- 直觉引入 → 背景知识 → 核心创新 → 实验验证 → 思考展望\n- 每个技术点都从\"为什么需要\"开始\n- 用类比连接新旧概念\n\n## 示例开头\n\n```\n你有没有想过这个问题：当我们问 ChatGPT \"中国的首都在哪里\"，它是怎么知道答案的？\n\n直觉上，我们可能会觉得模型\"记住\"了这个知识。但实际上，大模型根本没有这样的\"记忆模块\"——它所有的答案都是**算**出来的。\n\n这种方式有个明显的问题：**大炮打蚊子**。\n```\n\n## 适用场景\n\n技术博客、公众号文章、科普分享\n","content_type":"text/markdown; charset=utf-8","language":"markdown","size":1026,"content_sha256":"b66312f9d2a6be4f98cbab1935bfde7e8f7690002a0de2c0b956556d525881ba"},{"filename":"styles/with-code.md","content":"# with-code\n\n结合 GitHub 开源代码进行讲解\n\n## 使用方式\n\n- 克隆或浏览论文的 GitHub 仓库\n- 找到核心实现代码\n- 贴出关键代码片段并讲解\n- 将论文概念与代码实现对应\n\n## 示例\n\n```markdown\n论文中的门控机制在代码中是这样实现的：\n\n```python\n# engram/model.py\ndef compute_gate(self, hidden_state, memory_key):\n # 计算隐藏状态和记忆 key 的对齐分数\n score = torch.matmul(hidden_state, memory_key.T)\n gate = torch.sigmoid(score / self.temperature)\n return gate\n```\n\n可以看到，门控值就是隐藏状态和记忆 key 的点积，\n经过 sigmoid 归一化到 0-1 之间。\n```\n\n## 适用场景\n\n想要复现或深入理解实现细节的读者\n","content_type":"text/markdown; charset=utf-8","language":"markdown","size":735,"content_sha256":"fa3b15cf8e437eea51875efe80f4ce7ddd95a7db8da6ae85ee6c0316423a517c"},{"filename":"styles/with-formulas.md","content":"# with-formulas\n\n包含公式图片和详细讲解\n\n## 使用方式\n\n- 插入 MinerU 提取的公式图片\n- 对每个关键公式进行解读\n- 解释符号含义和计算过程\n- 说明公式的直觉意义\n\n## 示例\n\n```markdown\n![](images/formula_1.jpg)\n\n这个公式定义了记忆向量的构建方式：\n- $e_t$ 是最终的记忆向量\n- 不同 n-gram 长度和哈希头的结果被拼接\n- 这种设计让表容量可以独立扩展\n```\n\n## 适用场景\n\n需要深入理解数学细节的读者\n","content_type":"text/markdown; charset=utf-8","language":"markdown","size":505,"content_sha256":"bc374cbc68be5ae7a83b39219f8e6175475058106ae7ecf0dd7e7eb0bb4ad759"}],"content_json":{"type":"doc","content":[{"type":"heading","attrs":{"level":1},"content":[{"text":"Paper Analyzer — 学术论文深度解析","type":"text"}]},{"type":"paragraph","content":[{"text":"⚠️ ","type":"text"},{"text":"这是生产级指令。你的唯一任务：产出一篇让读者觉得\"比我读论文还清楚\"的深度HTML长文。","type":"text","marks":[{"type":"strong"}]}]},{"type":"heading","attrs":{"level":2},"content":[{"text":"快速使用","type":"text"}]},{"type":"code_block","attrs":{"wrap":false,"language":""},"content":[{"text":"/paper-analyzer https://arxiv.org/abs/2605.07363\n/paper-analyzer /path/to/paper.pdf\n/paper-analyzer 粘贴文本","type":"text"}]},{"type":"hr","attrs":{"markup":"---"}},{"type":"heading","attrs":{"level":2},"content":[{"text":"强制工作流（每一步必须执行，不可跳过）","type":"text"}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"Round 1：获取论文全文 ⛔","type":"text"}]},{"type":"table","attrs":{"layout":null},"content":[{"type":"tr","content":[{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"输入","type":"text"}]}]},{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"执行","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"arxiv URL","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"同时读","type":"text","marks":[{"type":"strong"}]},{"text":" arxiv.org/abs/（摘要）和 arxiv.org/html/（全文HTML）","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"PDF路径","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"用PDF读取工具读全文。分多次直到全部获取","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"文本","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"全部使用","type":"text"}]}]}]}]},{"type":"paragraph","content":[{"text":"自检","type":"text","marks":[{"type":"strong"}]},{"text":"：有没有完整内容？没有 → 换方式继续。","type":"text"}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"Round 2：搜索开源代码 ⛔","type":"text"}]},{"type":"ordered_list","attrs":{"order":1,"listStyle":"number"},"content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"从论文中提取代码仓库链接（通常在页脚或 Introduction 末）","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"没有则用论文标题+作者名搜索 GitHub","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"克隆：","type":"text"},{"text":"git clone --depth 1 \u003curl> /tmp/paper_code","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"阅读 README → 核心源码文件 → 配置文件","type":"text"}]}]}]},{"type":"paragraph","content":[{"text":"根据代码状态分支处理","type":"text","marks":[{"type":"strong"}]},{"text":"：","type":"text"}]},{"type":"table","attrs":{"layout":null},"content":[{"type":"tr","content":[{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"状态","type":"text"}]}]},{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"处理","type":"text"}]}]},{"type":"th","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"文章体现","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"✅ 已发布","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"读核心文件，找 ≥2 处论文方法↔源码对应","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"贴代码段（≤30行），标注 ","type":"text"},{"text":"文件路径:行号","type":"text","marks":[{"type":"code_inline"}]}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"⏳ 待发布","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"检查 README/Release 标记","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"标注状态+仓库链接","type":"text"}]}]}]},{"type":"tr","content":[{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"❌ 无代码","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"搜索替代实现/相关项目","type":"text"}]}]},{"type":"td","attrs":{"colspan":1,"rowspan":1,"colwidth":null,"alignment":""},"content":[{"type":"paragraph","content":[{"text":"注明\"本文未提供公开代码\"","type":"text"}]}]}]}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"Round 3：深度分析 ⛔ 内部完成，不展示过程","type":"text"}]},{"type":"ordered_list","attrs":{"order":1,"listStyle":"number"},"content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"核心创新：论文做了什么别人没做的？（1-3个，每个一句话提炼）","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"方法细节：输入→处理→输出→为什么更好（每个创新画清楚这条线）","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"关键实验：哪个结果最有说服力？为什么？","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"论文弱点：作者自述 + 你的判断","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"代码对应：每个 component 对应哪个文件/函数","type":"text"}]}]}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"Round 4：询问用户 ⛔","type":"text"}]},{"type":"paragraph","content":[{"text":"必须问风格选择，用户未回则默认 academic。","type":"text"}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"Round 5：写作输出HTML ⛔","type":"text"}]},{"type":"paragraph","content":[{"text":"按选定风格的要求写，输出完整HTML。模板见下文。","type":"text"}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"Round 6：自我审查 ⛔","type":"text"}]},{"type":"paragraph","content":[{"text":"逐项检查，不通过则修改直到通过。","type":"text"}]},{"type":"hr","attrs":{"markup":"---"}},{"type":"heading","attrs":{"level":2},"content":[{"text":"三风格详细要求","type":"text"}]},{"type":"hr","attrs":{"markup":"---"}},{"type":"heading","attrs":{"level":3},"content":[{"text":"storytelling（故事型）— 像一篇公众号爆文","type":"text"}]},{"type":"paragraph","content":[{"text":"硬标准","type":"text","marks":[{"type":"strong"}]},{"text":"：","type":"text"}]},{"type":"bullet_list","content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"字数 ≥ 3000","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"段落 ≥ 15","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"引用论文原文 ≥ 3 处","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"生动类比/比喻 ≥ 2 个","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"结尾金句 1 句","type":"text"}]}]}]},{"type":"paragraph","content":[{"text":"结构要求（按顺序，缺一不可）","type":"text","marks":[{"type":"strong"}]},{"text":"：","type":"text"}]},{"type":"code_block","attrs":{"wrap":false,"language":""},"content":[{"text":"1. 钩子开头（2-3段）\n — 反常识问题 / 引人共鸣的场景 / 让人\"等等再说一遍？\"的事实\n — 不要直接讲技术。先让读者好奇。\n\n2. \"为什么会这样\"（3-4段）\n — 解释现有方法的逻辑和它的瓶颈\n — 用简单例子说明\n — 让读者感到\"确实需要一种新方法\"\n\n3. 核心洞察（1-2段）\n — 论文最关键的那一句话发现\n — 用一句话说清楚 + 一个类比强化\n\n4. 方法详解（5-8段，全文最重点）\n — 分步骤展开：怎么做 → 为什么这样设计 → 和旧方法的关键区别\n — 每个步骤配一个类比\n — 引用论文原文（公式/算法描述）≥ 3 处\n — 用对比表呈现新旧方法差异\n\n5. 实验效果（3-4段）\n — 最重要的实验结果 + 数据解读\n — 不只是报数字，要解释\"这意味着什么\"\n — 用表格呈现关键对比数据\n\n6. 深层意义（2-3段）\n — 这个工作对行业意味着什么\n — 不止一个角度：技术意义、产业意义、方法学意义\n\n7. 局限（1-2段）\n — 作者自述的局限 + 你的判断\n\n8. 收束（1段）\n — 回到开头的场景/问题，形成闭环\n — 读者带着\"我懂了\"的感觉离开\n\n9. 金句\n — 一句话，让人能记住并转述","type":"text"}]},{"type":"paragraph","content":[{"text":"写法要求","type":"text","marks":[{"type":"strong"}]},{"text":"：","type":"text"}]},{"type":"bullet_list","content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"多用\"你\"和读者对话（\"你有没有想过\"\"你猜怎么着\"）","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"段落短，一段不超过 4 句话","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"技术词出现时要立刻给\"人话解释\"","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"数据要翻译成可感知的东西（\"15 斤荔枝\"而不只是\"15 斤\"）","type":"text"}]}]}]},{"type":"hr","attrs":{"markup":"---"}},{"type":"heading","attrs":{"level":3},"content":[{"text":"academic（学术型）— 比论文更清晰的深度解析","type":"text"}]},{"type":"paragraph","content":[{"text":"硬标准","type":"text","marks":[{"type":"strong"}]},{"text":"：","type":"text"}]},{"type":"bullet_list","content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"字数 ≥ 4000（⚠️ 学术型必须长于故事型）","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"段落 ≥ 20","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"论文公式引用 ≥ 5 处（用 KaTeX 渲染）","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"论文图片/图表引用 ≥ 3 处（标注 Figure number）","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"实验数据表格 ≥ 2 张","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"代码段 ≥ 2 段（如有代码）","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"指出局限 ≥ 2 处","type":"text"}]}]}]},{"type":"paragraph","content":[{"text":"结构要求","type":"text","marks":[{"type":"strong"}]},{"text":"：","type":"text"}]},{"type":"code_block","attrs":{"wrap":false,"language":""},"content":[{"text":"1. 论文元信息\n 标题 · 作者 · 链接 · 代码状态\n\n2. 一句话总结（100字内）\n\n3. 研究背景与动机（4-5段）\n — 这个领域在解决什么问题\n — 现有方法及其局限（按时间线或方法论分类）\n — 本文的出发点\n\n4. 预备知识（2-3段，如需要）\n — 理解本文需要的核心概念\n — 本文用到的基础方法简介\n\n5. 方法详解（8-10段，全文最重点）\n — 对每个创新点独立成节\n — 每个创新点包含：①问题 ②怎么做（配公式）③为什么有效 ④与已有方法的差异\n — 公式用 $...$ KaTeX 渲染\n — 引论文原文 Figure/Table 编号\n — 有代码则穿插源码分析\n\n6. 实验分析（4-6段）\n — 实验设置概述\n — 主要结果（配表格 + 深入解读）\n — 不同维度的对比分析\n — 消融实验说明了什么\n — 不是报数据，是解读数据背后的含义\n\n7. 讨论（2-3段）\n — 方法的适用边界\n — 未解决的问题\n — 对未来工作的启示\n\n8. 局限分析（2-3段）\n — 作者自述 ≥ 1 处\n — 你的独立判断 ≥ 1 处\n\n9. 结论（1-2段）\n — 凝练贡献\n — 展望","type":"text"}]},{"type":"paragraph","content":[{"text":"写法要求","type":"text","marks":[{"type":"strong"}]},{"text":"：","type":"text"}]},{"type":"bullet_list","content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"保持学术严谨但不死板——比论文好读","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"每个公式后要跟一句\"人话\"解释：这个公式在说什么","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"引用论文的 Fig/Table/Section 编号","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"表格数据要有解读，不只贴数据","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"数学符号首次出现要解释含义","type":"text"}]}]}]},{"type":"hr","attrs":{"markup":"---"}},{"type":"heading","attrs":{"level":3},"content":[{"text":"concise（精炼型）— 最快掌握核心","type":"text"}]},{"type":"paragraph","content":[{"text":"⚠️ ","type":"text"},{"text":"精炼 ≠ 敷衍。精炼是信息密度极高、但该有的全有。","type":"text","marks":[{"type":"strong"}]}]},{"type":"paragraph","content":[{"text":"硬标准","type":"text","marks":[{"type":"strong"}]},{"text":"：","type":"text"}]},{"type":"bullet_list","content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"字数 ≥ 1200（不能低于这个数）","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"必须有：核心摘要盒 + 表格 + 可视化图表 + 金句","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"⚠️ ","type":"text"},{"text":"必须包含至少 1 个 Mermaid 图表","type":"text","marks":[{"type":"strong"}]},{"text":"（架构图或对比图）","type":"text"}]}]}]},{"type":"paragraph","content":[{"text":"结构要求","type":"text","marks":[{"type":"strong"}]},{"text":"：","type":"text"}]},{"type":"code_block","attrs":{"wrap":false,"language":""},"content":[{"text":"1. 头图（Mermaid图表）—— 全文最核心架构/对比的一张图\n 类型可以是：flowchart（流程图）、graph（对比图）、或 timeline\n\n2. 核心摘要盒\n — 5 行以内\n — 覆盖：做什么 / 怎么做 / 效果 / 适用场景\n\n3. 关键创新（3-5 个，编号列出）\n — 每个 2-4 句\n — 一句话说创新点 → 一句话说怎么做的 → 一句话说为什么重要\n\n4. 核心数据表\n — 最多 5 行数据\n — 突出和 baseline 的对比\n\n5. 金句收尾","type":"text"}]},{"type":"paragraph","content":[{"text":"Mermaid 图表示例","type":"text","marks":[{"type":"strong"}]},{"text":"（⚠️ 节点文本避免中文特殊字符，用英文或简单ASCII。用 ","type":"text"},{"text":"\u003cbr/>","type":"text","marks":[{"type":"code_inline"}]},{"text":" 换行）：","type":"text"}]},{"type":"code_block","attrs":{"wrap":false,"language":"mermaid"},"content":[{"text":"flowchart TB\n subgraph DSA[\"DSA: 64 heads scan all L tokens\"]\n Q1[Query] --> H1[Head 1..64]\n H1 --> TK1[Score: O(64L)]\n end\n subgraph MISA[\"MISA: route to h=8 heads\"]\n Q2[Query] --> RTR[Router: O(64M)]\n RTR -->|top-8| H2[8 active heads]\n H2 --> TK2[Score: O(8L)]\n end\n DSA -->|8x fewer heads| MISA","type":"text"}]},{"type":"hr","attrs":{"markup":"---"}},{"type":"heading","attrs":{"level":2},"content":[{"text":"HTML 输出模板","type":"text"}]},{"type":"paragraph","content":[{"text":"生成HTML时使用此模板，确保含 KaTeX 公式渲染 + Mermaid 图表支持：","type":"text"}]},{"type":"code_block","attrs":{"wrap":false,"language":"html"},"content":[{"text":"\u003c!DOCTYPE html>\n\u003chtml lang=\"zh-CN\">\n\u003chead>\n\u003cmeta charset=\"UTF-8\">\n\u003cmeta name=\"viewport\" content=\"width=device-width, initial-scale=1.0\">\n\u003ctitle>论文标题 — 深度解读\u003c/title>\n\u003cstyle>\n:root{--text:#1a1a1a;--bg:#fafaf8;--accent:#2563eb;--muted:#6b7280;--border:#e5e7eb;--code-bg:#f3f4f6}\n*{margin:0;padding:0;box-sizing:border-box}\nbody{font-family:-apple-system,\"PingFang SC\",\"Noto Serif SC\",serif;color:var(--text);background:var(--bg);line-height:1.85;padding:2.5rem 1.5rem;max-width:720px;margin:0 auto;font-size:17px}\nh1{font-size:2rem;margin:0 0 .3rem;line-height:1.3}\nh2{font-size:1.35rem;margin:2.8rem 0 .8rem;color:var(--accent);padding-bottom:.4rem;border-bottom:1px solid var(--border)}\nh3{font-size:1.1rem;margin:1.5rem 0 .5rem;color:#333}\n.meta{color:var(--muted);font-size:.9rem;margin-bottom:2.5rem;line-height:1.8}\n.meta a{color:var(--accent);text-decoration:none}\nblockquote{border-left:3px solid var(--accent);padding:.6rem 1.2rem;margin:1.5rem 0;background:#f0f4ff;border-radius:0 8px 8px 0}\npre{background:var(--code-bg);padding:1rem 1.2rem;border-radius:8px;overflow-x:auto;font-size:.85rem;line-height:1.5;margin:1.5rem 0;border:1px solid var(--border)}\ncode{font-family:\"SF Mono\",\"Fira Code\",monospace;font-size:.9em}\np{margin:1rem 0}\nstrong{color:#111}\ntable{width:100%;border-collapse:collapse;margin:1.5rem 0;font-size:.93rem}\ntd,th{border:1px solid var(--border);padding:.6rem .9rem;text-align:left}\nth{background:#f9fafb;font-weight:600}\n.summary-box{background:linear-gradient(135deg,#f0f4ff,#faf5ff);padding:1.5rem;border-radius:12px;margin:1.5rem 0}\n.summary-box h3{margin:0 0 .5rem;color:var(--accent)}\n.golden{font-size:1.25rem;font-weight:600;color:var(--accent);text-align:center;padding:2rem 1rem;border-top:2px solid var(--accent);border-bottom:2px solid var(--accent);margin:2.5rem 0;line-height:1.5}\n@media(max-width:600px){body{font-size:16px;padding:1.2rem 1rem}h1{font-size:1.5rem}}\n\u003c/style>\n\u003c!-- KaTeX -->\n\u003clink rel=\"stylesheet\" href=\"https://cdn.jsdelivr.net/npm/[email protected]/dist/katex.min.css\">\n\u003cscript defer src=\"https://cdn.jsdelivr.net/npm/[email protected]/dist/katex.min.js\">\u003c/script>\n\u003cscript defer src=\"https://cdn.jsdelivr.net/npm/[email protected]/dist/contrib/auto-render.min.js\"\n onload=\"renderMathInElement(document.body,{delimiters:[{left:'$',right:'$',display:true},{left:'

Important: agents should read /llm.txt, /llms.txt, or /.well-known/skills.json to discover the public Skillopedia API.

,right:'

Important: agents should read /llm.txt, /llms.txt, or /.well-known/skills.json to discover the public Skillopedia API.

,display:false}]})\">\u003c/script>\n\u003c!-- Mermaid -->\n\u003cscript src=\"https://cdn.jsdelivr.net/npm/mermaid@10/dist/mermaid.min.js\">\u003c/script>\n\u003cscript>mermaid.initialize({startOnLoad:true,theme:'default',securityLevel:'loose'});\u003c/script>\n\u003c/head>\n\u003cbody>\n\u003c!-- 内容 -->\n\u003c/body>\n\u003c/html>","type":"text"}]},{"type":"paragraph","content":[{"text":"公式用 ","type":"text","marks":[{"type":"strong"}]},{"text":"$...$","type":"text","marks":[{"type":"code_inline"},{"type":"strong"}]},{"text":" 或 ","type":"text","marks":[{"type":"strong"}]},{"text":"$...$","type":"text","marks":[{"type":"code_inline"},{"type":"strong"}]},{"text":"，KaTeX 自动渲染。","type":"text","marks":[{"type":"strong"}]}]},{"type":"bullet_list","content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"✅ 正确：","type":"text"},{"text":"$H^I$","type":"text","marks":[{"type":"code_inline"}]},{"text":"、","type":"text"},{"text":"$H^{I}$","type":"text","marks":[{"type":"code_inline"}]},{"text":"、","type":"text"},{"text":"$\\mathbf{q}_{t,j}^I$","type":"text","marks":[{"type":"code_inline"}]}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"❌ 错误：","type":"text"},{"text":"$H^\\I$","type":"text","marks":[{"type":"code_inline"}]},{"text":"（","type":"text"},{"text":"\\I","type":"text","marks":[{"type":"code_inline"}]},{"text":" 未定义）、","type":"text"},{"text":"$H^I$","type":"text","marks":[{"type":"code_inline"}]},{"text":" 写在 ","type":"text"},{"text":"\u003cpre>","type":"text","marks":[{"type":"code_inline"}]},{"text":" 标签内","type":"text"}]}]}]},{"type":"paragraph","content":[{"text":"Mermaid 图用 ","type":"text","marks":[{"type":"strong"}]},{"text":"\u003cpre class=\"mermaid\">...\u003c/pre>","type":"text","marks":[{"type":"code_inline"},{"type":"strong"}]},{"text":" 包裹。节点文本避免中文标点和特殊字符。","type":"text","marks":[{"type":"strong"}]}]},{"type":"hr","attrs":{"markup":"---"}},{"type":"heading","attrs":{"level":2},"content":[{"text":"自我审查清单（Round 6）","type":"text"}]},{"type":"paragraph","content":[{"text":"生成后逐条检查，不通过则修改：","type":"text"}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"通用","type":"text"}]},{"type":"checkbox_list","attrs":{"id":null},"content":[{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"字数达标？（story≥3000 / academic≥4000 / concise≥1200）","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"引用论文原文 ≥ 3 处？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"每个核心创新独立深度展开？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"至少 1 个实验结果做深入解读？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"代码状态已提及？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"有代码则源码 ≥ 2 段 + 文件路径？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"指出局限 ≥ 2 处（至少 1 处是作者自述的）？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"HTML 格式完整，可在浏览器打开？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"无 AI 套话（\"深入探讨\"\"至关重要\"\"值得注意的是\"）？","type":"text"}]}]}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"storytelling 专属","type":"text"}]},{"type":"checkbox_list","attrs":{"id":null},"content":[{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"有钩子开头？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"有 ≥ 2 个类比/比喻？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"用\"你\"和读者对话？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"有收束段落形成闭环？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"有金句？","type":"text"}]}]}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"academic 专属","type":"text"}]},{"type":"checkbox_list","attrs":{"id":null},"content":[{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"字数 ≥ storytelling？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"公式 ≥ 5 处（KaTeX 渲染）？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"论文图/表引用 ≥ 3 处（Fig/Table 编号）？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"实验数据表 ≥ 2 张？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"方法部分 ≥ 8 段？","type":"text"}]}]}]},{"type":"heading","attrs":{"level":3},"content":[{"text":"concise 专属","type":"text"}]},{"type":"checkbox_list","attrs":{"id":null},"content":[{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"有 Mermaid 图表？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"有核心摘要盒？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"有对比数据表？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"有金句？","type":"text"}]}]},{"type":"checkbox_item","attrs":{"checked":false},"content":[{"type":"paragraph","content":[{"text":"字数 ≥ 1200？","type":"text"}]}]}]},{"type":"hr","attrs":{"markup":"---"}},{"type":"heading","attrs":{"level":2},"content":[{"text":"参考文件","type":"text"}]},{"type":"bullet_list","content":[{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"styles/storytelling.md","type":"text","marks":[{"type":"code_inline"}]},{"text":" — 故事型补充规范","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"styles/academic.md","type":"text","marks":[{"type":"code_inline"}]},{"text":" — 学术型补充规范","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"styles/concise.md","type":"text","marks":[{"type":"code_inline"}]},{"text":" — 精炼型补充规范","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"styles/with-formulas.md","type":"text","marks":[{"type":"code_inline"}]},{"text":" — 公式详解","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"styles/with-code.md","type":"text","marks":[{"type":"code_inline"}]},{"text":" — 代码分析规范","type":"text"}]}]},{"type":"list_item","content":[{"type":"paragraph","content":[{"text":"scripts/generate_html.py","type":"text","marks":[{"type":"code_inline"}]},{"text":" — HTML生成辅助脚本","type":"text"}]}]}]},{"type":"hr","attrs":{"markup":"---"}}]},"metadata":{"date":"2026-06-05","name":"paper-analyzer","author":"@skillopedia","source":{"stars":389,"repo_name":"paper-craft-skills","origin_url":"https://github.com/zsyggg/paper-craft-skills/blob/HEAD/skills/paper-analyzer/SKILL.md","repo_owner":"zsyggg","body_sha256":"a30bab91150d9c8f5d45b8d6fc9f4edf0c9999bf424ce15f6247a89091de2c51","cluster_key":"f394d3dfaa2a94065e7e81900053ca3c519a4da2794b2bb2a2eb38515c449d2b","clean_bundle":{"format":"clean-skill-bundle-v1","source":"zsyggg/paper-craft-skills/skills/paper-analyzer/SKILL.md","attachments":[{"id":"87049ed4-c9ab-5f0d-a0d9-45a5a9fc2eb5","key":"uploads/10433ee7-ad12-4ae0-b34e-97553e46c6c8/87049ed4-c9ab-5f0d-a0d9-45a5a9fc2eb5/attachment.txt","path":"cover-prompt-simple.txt","size":767,"sha256":"95a55a5e59bb3e5c02932c6308b804ed0c62877cdb3a2971898dc95608f74d48","contentType":"text/plain; charset=utf-8"},{"id":"6205f126-cc84-5ad7-9246-a41c4c325f4a","key":"uploads/10433ee7-ad12-4ae0-b34e-97553e46c6c8/6205f126-cc84-5ad7-9246-a41c4c325f4a/attachment.md","path":"cover-prompt.md","size":1068,"sha256":"4d85f7224391976ed260e34eb985ac64e34a9e5b0e50ddda4ce340aedcc60473","contentType":"text/markdown; charset=utf-8"},{"id":"6fba0c2b-e724-5d1b-a490-a86b8290b67f","key":"uploads/10433ee7-ad12-4ae0-b34e-97553e46c6c8/6fba0c2b-e724-5d1b-a490-a86b8290b67f/attachment.py","path":"scripts/extract_paper_info.py","size":3669,"sha256":"4850096c02a10ea9a6da8aaa0dbd5d699db866784eb2342bb142204ed665ab85","contentType":"text/x-python; charset=utf-8"},{"id":"71e8b9ba-0940-5800-91ea-7a7f7aa5b99b","key":"uploads/10433ee7-ad12-4ae0-b34e-97553e46c6c8/71e8b9ba-0940-5800-91ea-7a7f7aa5b99b/attachment.py","path":"scripts/generate_html.py","size":3963,"sha256":"396200c5eafe58b417715a778382486d77a5532c4aeca1c9d94d97e7a0162e38","contentType":"text/x-python; charset=utf-8"},{"id":"a524467c-8977-5a63-adae-8ff86363c466","key":"uploads/10433ee7-ad12-4ae0-b34e-97553e46c6c8/a524467c-8977-5a63-adae-8ff86363c466/attachment.md","path":"styles/academic.md","size":1054,"sha256":"e2c33a3a63921e6e716da4aedef6ce1d5256ff0e36dda4d141149509a0440f05","contentType":"text/markdown; charset=utf-8"},{"id":"6335289f-d81c-5cd6-b10e-f3b77a73a59b","key":"uploads/10433ee7-ad12-4ae0-b34e-97553e46c6c8/6335289f-d81c-5cd6-b10e-f3b77a73a59b/attachment.md","path":"styles/concise.md","size":949,"sha256":"285ad5029e537f452010b8b002a6873ed67121001e8c8c2d3573edaec5e5fc33","contentType":"text/markdown; charset=utf-8"},{"id":"c4fa8588-6212-5096-9656-dd67819b9607","key":"uploads/10433ee7-ad12-4ae0-b34e-97553e46c6c8/c4fa8588-6212-5096-9656-dd67819b9607/attachment.md","path":"styles/storytelling.md","size":1026,"sha256":"b66312f9d2a6be4f98cbab1935bfde7e8f7690002a0de2c0b956556d525881ba","contentType":"text/markdown; charset=utf-8"},{"id":"1fa733a2-8170-526d-8492-393bc415f62e","key":"uploads/10433ee7-ad12-4ae0-b34e-97553e46c6c8/1fa733a2-8170-526d-8492-393bc415f62e/attachment.md","path":"styles/with-code.md","size":735,"sha256":"fa3b15cf8e437eea51875efe80f4ce7ddd95a7db8da6ae85ee6c0316423a517c","contentType":"text/markdown; charset=utf-8"},{"id":"ec860d76-d454-5844-8224-086e439a8489","key":"uploads/10433ee7-ad12-4ae0-b34e-97553e46c6c8/ec860d76-d454-5844-8224-086e439a8489/attachment.md","path":"styles/with-formulas.md","size":505,"sha256":"bc374cbc68be5ae7a83b39219f8e6175475058106ae7ecf0dd7e7eb0bb4ad759","contentType":"text/markdown; charset=utf-8"}],"bundle_sha256":"7059441f26f0103307b2ef148b265f105ef79afb2f49d95b6d52d0fff40fcd0d","attachment_count":9,"text_attachments":9,"attachment_storage":"skillopedia-attachments-v1","binary_attachments":0,"excluded_attachments":[]},"cluster_size":1,"skill_md_path":"skills/paper-analyzer/SKILL.md","import_metadata":{"date":"2026-06-05","author":"@skillopedia","version":"v1","category":"web-development","category_label":"Web"},"exact_dupes_collapsed_into_this":0},"version":"v1","category":"web-development","import_tag":"clean-skills-v1","description":"将学术论文转化为深度HTML长文。6轮强制工作流、代码仓库搜索、公式渲染、Mermaid图表。\n3种写作风格，输出可直接分享的精美HTML页面。\n"}},"renderedAt":1782980318588}

Important: agents should read /llm.txt, /llms.txt, or /.well-known/skills.json to discover the public Skillopedia API.