import live www.sillylaird.ca webrootHEADmain

6 files changed, 819 insertions, 0 deletions

diff --git a/.agents/skills/daily-paper-generator/SKILL.md b/.agents/skills/daily-paper-generator/SKILL.md
new file mode 100644
index 0000000..bdc8254
--- /dev/null
+++ b/.agents/skills/daily-paper-generator/SKILL.md
@@ -0,0 +1,273 @@
+---
+name: daily-paper-generator
+description: Use when the user asks to "generate daily paper", "search arXiv for EEG papers", "find EEG decoding papers", "review brain-computer interface papers", or wants to create paper summaries for EEG/brain decoding/speech decoding research. This skill automates searching arXiv for recent papers on EEG decoding, EEG speech decoding, or brain foundation models, reviewing paper quality, and generating structured Chinese/English summaries.
+version: 0.4.0
+---
+
+# Daily Paper Generator
+
+## Overview
+
+Automate the workflow of discovering, reviewing, and summarizing recent research papers on arXiv related to EEG decoding, brain-computer interfaces, and neural foundation models.
+
+**Core workflow:**
+1. Search arXiv for recent papers (within 3 months) using Chrome browser
+2. Retrieve paper metadata from arXiv pages
+3. Evaluate paper quality using structured criteria
+4. Select top 3 papers
+5. Generate structured summaries with Chinese and English reviews
+6. Save results as Markdown files in `daily paper/` directory
+
+## When to Use
+
+Use this skill when:
+- User asks to "generate daily paper" or "find recent EEG papers"
+- User wants to discover research on EEG decoding, speech decoding from EEG, or brain foundation models
+- User needs paper reviews with both Chinese and English summaries
+- User wants to track recent arXiv publications in neuro/AI intersection
+
+## Output Format
+
+Each paper summary follows this structure (see `example/daily paper example.md` for complete example):
+
+### 1. Header Section
+```markdown
+# Paper Title
+
+## 作者及单位
+Author list
+Institution
+
+## arXiv 链接
+https://arxiv.org/abs/ARXIV_ID
+
+**发表日期**: YYYY-MM-DD
+**arXiv ID**: XXXX.XXXXX
+**分类**: cs.LG, q-bio.NC, eess.SP
+```
+
+### 2. Review Sections
+
+**中文评语** (~300 words):
+- Background (1-2 sentences): Research context and importance
+- Challenges (2-3 sentences): Problems with existing methods
+- Contribution (1-2 sentences): Core contribution of this work
+- Method (2-3 sentences): Key technical details
+- Results (2-3 sentences): Main findings and metrics
+- Analysis & Limitations (1-2 sentences): Significance and limitations
+
+**English Review** (fluent academic English):
+- Concise summary following the same structure as Chinese review
+- Use natural academic prose (avoid AI-like patterns)
+- Apply scientific writing best practices
+
+### 3. Main Figure Section
+```markdown
+## 主图
+[预留论文主图位置]
+```
+
+### 4. Metadata Table
+```markdown
+## 论文元数据
+
+| 项目 | 内容 |
+|------|------|
+| **标题** | Paper Title |
+| **第一作者** | First Author Name |
+| **作者列表** | Full author list |
+| **第一作者单位** | Institution |
+| **发表日期** | YYYY-MM-DD |
+| **arXiv 链接** | https://arxiv.org/abs/ID |
+| **PDF 链接** | https://arxiv.org/pdf/ID |
+| **分类** | cs.LG, q-bio.NC, eess.SP |
+```
+
+### 5. Integrated Format (for publishing)
+```markdown
+## 整合格式
+
+Daily Paper MMDD
+
+Paper Title
+
+https://arxiv.org/abs/ARXIV_ID
+
+[Chinese Review]
+
+[English Review]
+```
+
+### 6. Appendix
+```markdown
+## 附录
+
+**github连接：** [Available/Not Available]
+
+**补充说明**
+
+[Key insights, impact points]
+
+**Sources:**
+- [arXiv Abstract](URL)
+- [arXiv HTML](URL)
+- [Paperverse Review](URL) (if available)
+```
+
+## Quick Reference
+
+| Task | Method |
+|------|--------|
+| Search arXiv | Use Chrome MCP tools (chrome-mcp-helper) |
+| Get paper details | Navigate to arXiv pages and extract metadata |
+| Evaluate quality | Use criteria in `references/quality-criteria.md` |
+| Write Chinese review | Follow style in `references/writing-style.md` |
+| Write English review | Apply scientific-writing skill best practices |
+| Create output | Use template in `example/daily paper example.md` |
+
+## Workflow
+
+### Step 1: Search arXiv Using Chrome
+
+**Search keywords** (see `references/keywords.md` for full list):
+- EEG decoding: `EEG decoding`, `brain decoding`, `neural decoding`
+- Speech decoding: `speech decoding from EEG`, `EEG speech reconstruction`
+- Foundation models: `EEG foundation model`, `large EEG model`, `brain foundation model`
+
+**Method: Use Chrome browser with arXiv search**
+
+1. **Navigate to arXiv search** using Chrome MCP tools:
+   - URL: `https://arxiv.org/search/`
+   - Add search parameters: `?searchtype=all&query=KEYWORDS&abstracts=show&order=-announced_date_first`
+
+2. **Search URL pattern**:
+   ```
+   https://arxiv.org/search/?searchtype=all&query=EEG+decoding&abstracts=show&order=-announced_date_first
+   https://arxiv.org/search/?searchtype=all&query=EEG+foundation+model&abstracts=show&order=-announced_date_first
+   ```
+
+3. **Time filtering**: Use date filters or sort by `announced_date_first` to get recent papers
+
+4. **Extract paper information** from search results:
+   - Paper title
+   - Authors
+   - arXiv ID
+   - Abstract preview
+   - Publication date
+
+### Step 2: Retrieve Paper Details
+
+For each candidate paper, navigate to its arXiv abs page and extract:
+
+**URL pattern**: `https://arxiv.org/abs/ARXIV_ID`
+
+**Extract from page**:
+- Title (from `<h1>` tag)
+- Authors (from `.authors` element)
+- Abstract (from `blockquote.abstract`)
+- Submission date (from `.dateline`)
+- arXiv ID (from URL or page)
+- Categories (from `.subjects`)
+- Comments (if present)
+- First author institution (if available in comments or author affiliations)
+
+### Step 3: Evaluate Paper Quality
+
+Review each paper using the 5-dimension criteria in `references/quality-criteria.md`:
+
+| Dimension | Weight | Key Points |
+|-----------|--------|------------|
+| Innovation | 30% | Novelty of contribution |
+| Method Completeness | 25% | Clarity and reproducibility |
+| Experimental Thoroughness | 25% | Validation depth |
+| Writing Quality | 10% | Clarity of expression |
+| Relevance & Impact | 10% | Domain importance |
+
+**Scoring:** Rate each dimension 1-5, calculate weighted sum.
+
+**Process:**
+1. Screen by title/abstract for relevance
+2. Navigate to full paper page for detailed review
+3. Score each dimension
+4. Rank by total score
+5. Select top 3
+
+### Step 4: Generate Paper Summaries
+
+For each selected paper, create a summary following the structure in `example/daily paper example.md`:
+
+**Required sections:**
+1. Title (H1 heading)
+2. 作者及单位 (Authors and Institution)
+3. arXiv 链接 (with metadata: date, ID, categories)
+4. 中文评语 (Chinese review, ~300 words)
+5. English Review (fluent academic English)
+6. 主图 (placeholder for main figure)
+7. 论文元数据 (metadata table)
+8. 整合格式 (integrated format for publishing)
+9. 附录 (appendix with github link,补充说明, sources)
+
+**Writing Chinese review** (see `references/writing-style.md`):
+- Background: 研究背景和重要性
+- Challenges: 现有方法的不足
+- Contribution: 本工作的核心贡献
+- Method: 关键技术细节
+- Results: 主要发现和指标
+- Analysis & Limitations: 意义和局限性
+
+**Writing English review**:
+- Apply scientific-writing skill best practices
+- Use anti-AI writing principles (natural, varied sentence structure)
+- Keep concise and direct
+- Avoid formulaic transitions ("furthermore", "moreover", "additionally")
+
+### Step 5: Save Output
+
+Create Markdown files in the `daily paper/` directory:
+
+```
+daily paper/
+├── 2025-01-26-1430-paper-1.md
+├── 2025-01-26-1430-paper-2.md
+└── 2025-01-26-1430-paper-3.md
+```
+
+**Filename format:** `YYYY-MM-DD-HHMM-paper-N.md`
+
+**Important:** 使用时间戳（精确到分钟）避免覆盖之前生成的文件。
+
+## Example Output
+
+See `example/daily paper example.md` for a complete example of the DeeperBrain paper summary with all sections properly formatted.
+
+## Additional Resources
+
+### Reference Files
+
+- **`references/keywords.md`** - Complete search keyword list and arXiv URL patterns
+- **`references/quality-criteria.md`** - Detailed 5-dimension evaluation criteria with scoring rubrics
+- **`references/writing-style.md`** - Chinese review structure, templates, and example analysis
+
+### Example Files
+
+- **`example/daily paper example.md`** - Complete output example with all sections
+- **`scripts/arxiv_search.py`** - Legacy Python script (deprecated, use Chrome instead)
+
+### Chrome MCP Tools
+
+Use Chrome MCP tools for browser automation:
+- **Navigation**: Open arXiv search and paper pages
+- **Screenshot**: Capture pages for analysis
+- **Tabs**: Manage multiple arXiv pages
+- **Content extraction**: Parse paper metadata from HTML
+
+## Important Notes
+
+1. **Time range:** Search focuses on papers from the last 3 months (check submission dates)
+2. **Link format:** Use arXiv abs page links (https://arxiv.org/abs/ID), not direct PDF links
+3. **Review length:** Chinese reviews should be approximately 300 words
+4. **Quality focus:** Prioritize content quality (innovation, method, experiments) over quantitative metrics
+5. **Bilingual output:** Both Chinese and English reviews are required for each paper
+6. **Chrome required:** This workflow uses Chrome browser automation via MCP tools
+7. **Complete format:** Ensure all 9 sections are included in each summary
+8. **Consistent naming:** Use Daily Paper MMDD format in integrated section
diff --git a/.agents/skills/daily-paper-generator/example/daily paper example.md b/.agents/skills/daily-paper-generator/example/daily paper example.md
new file mode 100644
index 0000000..8d95ede
--- /dev/null
+++ b/.agents/skills/daily-paper-generator/example/daily paper example.md
@@ -0,0 +1,78 @@
+# DeeperBrain: A Neuro-Grounded EEG Foundation Model Towards Universal BCI
+
+## 作者及单位
+Jiquan Wang, Sha Zhao, Yangxuan Zhou, Yiming Kang, Shijian Li, Gang Pan
+Zhejiang University, College of Computer Science and Technology
+
+## arXiv 链接
+https://arxiv.org/abs/2601.06134
+
+**发表日期**: 2026-01-05
+**arXiv ID**: 2601.06134
+**分类**: cs.LG, q-bio.NC, eess.SP
+
+---
+
+## 中文评语
+
+通用脑机接口的发展受限于 EEG 信号的跨受试和跨任务泛化能力不足。现有基础模型大多采用通用深度学习架构，忽略了 EEG 信号的神经生理学特性和生物物理约束，且在冻结探针评估下效果有限。本研究提出 DeeperBrain，一种神经驱动的 EEG 基础模型，将领域特定的归纳偏差整合到模型设计和学习目标中。在架构层面，该方法包含基于容积传导的通道编码和神经动力学感知的时间编码。在预训练层面，引入双目标策略：掩码 EEG 重建保证局部保真度，神经动力学统计预测以强制与宏观脑状态对齐。实验结果显示，DeeperBrain 在零样本跨受试迁移、跨任务泛化和少样本学习场景下优于现有基础模型。更重要的是，它在严格的冻结探针评估下保持优越效果，验证了将神经科学第一原理嵌入模型能够赋予学习表示通用 BCI 所需的泛化能力。
+
+## English Review
+
+Universal Brain-Computer Interfaces are constrained by the limited cross-subject and cross-task generalization of EEG signals. Most existing foundation models employ generic deep learning architectures that overlook neurophysiological characteristics and biophysical constraints, showing limited efficacy under frozen probing evaluation. This study presents DeeperBrain, a neuro-grounded EEG foundation model that integrates domain-specific inductive biases into both architecture design and learning objectives. The model incorporates volume conduction-based channel encoding and neurodynamics-aware temporal encoding to capture spatial and temporal patterns respectively. For pretraining, this study introduces a dual-objective strategy combining masked EEG reconstruction for local fidelity and neurodynamics statistics prediction to align with macroscopic brain states. Experiments show DeeperBrain outperforms existing foundation models across zero-shot cross-subject transfer, cross-task generalization, and few-shot learning scenarios. The model maintains strong performance under frozen probing evaluation, demonstrating that embedding neuroscientific first principles endows learned representations with the generalization needed for universal BCI.
+
+## 主图
+
+这里需要将论文的主图进行下载并存放。
+
+---
+
+## 论文元数据
+
+| 项目 | 内容 |
+|------|------|
+| **标题** | DeeperBrain: A Neuro-Grounded EEG Foundation Model Towards Universal BCI |
+| **第一作者** | Jiquan Wang |
+| **作者列表** | Jiquan Wang, Sha Zhao, Yangxuan Zhou, Yiming Kang, Shijian Li, Gang Pan |
+| **第一作者单位** | Zhejiang University, College of Computer Science and Technology |
+| **发表日期** | 2026-01-05 |
+| **arXiv 链接** | https://arxiv.org/abs/2601.06134 |
+| **PDF 链接** | https://arxiv.org/pdf/2601.06134 |
+| **分类** | cs.LG, q-bio.NC, eess.SP |
+
+---
+
+## 整合格式
+
+Daily Paper 0126
+
+DeeperBrain: A Neuro-Grounded EEG Foundation Model Towards Universal BCI
+
+https://arxiv.org/abs/2601.06134
+
+
+
+通用脑机接口的发展受限于 EEG 信号的跨受试和跨任务泛化能力不足。现有基础模型大多采用通用深度学习架构，忽略了 EEG 信号的神经生理学特性和生物物理约束，且在冻结探针评估下效果有限。本研究提出 DeeperBrain，一种神经驱动的 EEG 基础模型，将领域特定的归纳偏差整合到模型设计和学习目标中。在架构层面，该方法包含基于容积传导的通道编码和神经动力学感知的时间编码。在预训练层面，引入双目标策略：掩码 EEG 重建保证局部保真度，神经动力学统计预测以强制与宏观脑状态对齐。实验结果显示，DeeperBrain 在零样本跨受试迁移、跨任务泛化和少样本学习场景下优于现有基础模型。更重要的是，它在严格的冻结探针评估下保持优越效果，验证了将神经科学第一原理嵌入模型能够赋予学习表示通用 BCI 所需的泛化能力。
+
+
+
+Universal Brain-Computer Interfaces are constrained by the limited cross-subject and cross-task generalization of EEG signals. Most existing foundation models employ generic deep learning architectures that overlook neurophysiological characteristics and biophysical constraints, showing limited efficacy under frozen probing evaluation. This study presents DeeperBrain, a neuro-grounded EEG foundation model that integrates domain-specific inductive biases into both architecture design and learning objectives. The model incorporates volume conduction-based channel encoding and neurodynamics-aware temporal encoding to capture spatial and temporal patterns respectively. For pretraining, this study introduces a dual-objective strategy combining masked EEG reconstruction for local fidelity and neurodynamics statistics prediction to align with macroscopic brain states. Experiments show DeeperBrain outperforms existing foundation models across zero-shot cross-subject transfer, cross-task generalization, and few-shot learning scenarios. The model maintains strong performance under frozen probing evaluation, demonstrating that embedding neuroscientific first principles endows learned representations with the generalization needed for universal BCI.
+
+## 附录
+
+**github连接：**未开源
+
+**补充说明**
+
+这篇论文的重要价值在于：
+
+1. **跨学科融合**：将神经科学知识与深度学习结合
+2. **可解释性**：神经驱动设计提高了模型的可解释性
+3. **泛化能力**：在跨受试、跨任务场景下表现优异
+4. **实用价值**：为通用 BCI 系统的开发提供了新方向
+
+**Sources:**
+
+- [arXiv Abstract](https://arxiv.org/abs/2601.06134)
+- [arXiv HTML](https://arxiv.org/html/2601.06134v1)
+- [Paperverse Review](https://paperverse.io/paper/eabc5d58-8762-4dc9-aaf3-665057852cb7)
diff --git a/.agents/skills/daily-paper-generator/references/keywords.md b/.agents/skills/daily-paper-generator/references/keywords.md
new file mode 100644
index 0000000..10eccc3
--- /dev/null
+++ b/.agents/skills/daily-paper-generator/references/keywords.md
@@ -0,0 +1,69 @@
+# 搜索关键词列表
+
+## 核心搜索关键词
+
+### 脑电解码 (EEG Decoding)
+- `EEG decoding`
+- `brain decoding`
+- `neural decoding`
+- `EEG classification`
+- `brain signal decoding`
+
+### 脑电语音解码 (EEG Speech Decoding)
+- `speech decoding from EEG`
+- `EEG speech reconstruction`
+- `articulatory feature from EEG`
+- `EEG to speech`
+- `neural speech decoding`
+
+### 脑电大模型 (EEG Foundation Models)
+- `EEG foundation model`
+- `large EEG model`
+- `pretrained EEG model`
+- `brain foundation model`
+- `EEG representation learning`
+- `self-supervised EEG`
+- `contrastive learning EEG`
+
+### 相关主题 (Related Topics)
+- `brain-computer interface`
+- `BCI`
+- `EEG transformer`
+- `EEG language model`
+
+## arXiv 分类筛选
+
+推荐搜索以下 arXiv 分类：
+- `cs.CV` (Computer Vision and Pattern Recognition)
+- `cs.LG` (Machine Learning)
+- `cs.NE` (Neural and Evolutionary Computing)
+- `cs.AI` (Artificial Intelligence)
+- `cs.HC` (Human-Computer Interaction)
+- `q-bio.NC` (Neurons and Cognition)
+- `stat.ML` (Machine Learning)
+
+## 搜索策略
+
+1. **组合搜索**：将关键词组合以提高精度
+   - 例如：`EEG + speech + decoding`
+   - 例如：`foundation + model + EEG`
+
+2. **时间过滤**：搜索最近3个月的论文
+   - 使用 `submittedDate` 排序
+   - 手动过滤日期范围
+
+3. **分类过滤**：限制在相关分类
+   - 优先 `cs.CV`, `cs.LG`, `q-bio.NC`
+
+## arXiv API 查询示例
+
+```
+# 基础查询
+http://export.arxiv.org/api/query?search_query=all:EEG+decoding&start=0&max_results=50&sortBy=submittedDate&sortOrder=descending
+
+# 分类过滤查询
+http://export.arxiv.org/api/query?search_query=cat:cs.CV+OR+cat:cs.LG+OR+cat:q-bio.NC+AND+all:EEG+AND+speech+decoding&start=0&max_results=50&sortBy=submittedDate&sortOrder=descending
+
+# 大模型查询
+http://export.arxiv.org/api/query?search_query=all:EEG+AND+foundation+model&start=0&max_results=50&sortBy=submittedDate&sortOrder=descending
+```
diff --git a/.agents/skills/daily-paper-generator/references/quality-criteria.md b/.agents/skills/daily-paper-generator/references/quality-criteria.md
new file mode 100644
index 0000000..bb360bd
--- /dev/null
+++ b/.agents/skills/daily-paper-generator/references/quality-criteria.md
@@ -0,0 +1,113 @@
+# 论文质量评审标准
+
+## 评审维度与权重
+
+| 维度 | 权重 | 说明 |
+|------|------|------|
+| **创新性** | 30% | 论文的创新程度和贡献的新颖性 |
+| **方法完整性** | 25% | 方法的描述完整性和可复现性 |
+| **实验充分性** | 25% | 实验设计的全面性和结果的可信度 |
+| **写作质量** | 10% | 论文表达的清晰度和学术规范性 |
+| **相关性与影响力** | 10% | 与领域的相关性和潜在影响力 |
+
+## 详细评分标准
+
+### 1. 创新性 (30%)
+
+| 分数 | 标准 |
+|------|------|
+| **5分 - 突破性贡献** | 提出全新的范式或方法，对领域有重大影响 |
+| **4分 - 显著创新** | 在现有方法上有显著改进，提出新的见解 |
+| **3分 - 方法创新** | 提出了新的方法或框架，有一定的创新性 |
+| **2分 - 改进型** | 对现有方法有改进，但创新有限 |
+| **1分 - 增量改进** | 仅有微小的改进或组合现有方法 |
+
+**评估要点：**
+- 是否提出了新的问题或视角？
+- 方法是否有实质性创新？
+- 是否突破了现有方法的局限？
+
+### 2. 方法完整性 (25%)
+
+| 分数 | 标准 |
+|------|------|
+| **5分 - 完整且严谨** | 方法描述完整，数学推导严谨，易于复现 |
+| **4分 - 非常完整** | 方法描述详细，大部分细节可复现 |
+| **3分 - 可复现** | 核心方法清晰，可基本复现 |
+| **2分 - 缺乏细节** | 关键细节缺失，复现困难 |
+| **1分 - 表述不清** | 方法描述不清楚，无法判断有效性 |
+
+**评估要点：**
+- 方法描述是否清晰？
+- 是否提供了足够的细节？
+- 是否有代码仓库？
+- 数学推导是否严谨？
+
+### 3. 实验充分性 (25%)
+
+| 分数 | 标准 |
+|------|------|
+| **5分 - 全面深入** | 多数据集验证，充分消融实验，详细分析 |
+| **4分 - 非常充分** | 多个数据集，合理消融实验 |
+| **3分 - 合理验证** | 主干实验完整，结果可信 |
+| **2分 - 验证不足** | 实验较少，缺乏对比 |
+| **1分 - 实验不足** | 仅在简单场景验证，结果不可信 |
+
+**评估要点：**
+- 是否在标准数据集上验证？
+- 是否有充分的对比实验？
+- 是否有消融实验？
+- 统计显著性如何？
+
+### 4. 写作质量 (10%)
+
+| 分数 | 标准 |
+|------|------|
+| **5分 - 优秀** | 表达清晰，逻辑严密，学术规范 |
+| **4分 - 良好** | 表达清楚，逻辑基本完整 |
+| **3分 - 清晰** - 表达基本清晰，可理解 |
+| **2分 - 一般** | 表述有模糊之处 |
+| **1分 - 表述不清** | 表述混乱，难以理解 |
+
+### 5. 相关性与影响力 (10%)
+
+| 分数 | 标准 |
+|------|------|
+| **5分 - 广泛影响** | 解决重要问题，影响多个领域 |
+| **4分 - 领域重要** | 解决领域内重要问题 |
+| **3分 - 相关有意义** - 研究有意义，有一定影响 |
+| **2分 - 小众问题** | 针对小众问题 |
+| **1分 - 影响有限** - 影响非常有限 |
+
+## 自动评分辅助指标
+
+在人工评审前，可使用以下指标辅助初筛：
+
+- **摘要质量**：摘要是否包含实验结果和具体数据
+- **数据集**：是否在知名数据集上验证（如 TUH EEG, BCIC IV 等）
+- **代码可用性**：是否提供 GitHub 链接
+- **作者机构**：第一作者/机构的学术声誉（可选）
+- **引用数**：arXiv 上的早期引用数（可选）
+
+## 综合评分计算
+
+```
+总分 = 创新性×0.30 + 方法完整性×0.25 + 实验充分性×0.25 + 写作质量×0.10 + 相关性与影响力×0.10
+```
+
+**评分示例：**
+- 创新性：4分
+- 方法完整性：3分
+- 实验充分性：4分
+- 写作质量：3分
+- 相关性与影响力：4分
+
+总分 = 4×0.30 + 3×0.25 + 4×0.25 + 3×0.10 + 4×0.10 = 1.2 + 0.75 + 1.0 + 0.3 + 0.4 = 3.65
+
+## 评审流程
+
+1. **初筛**：根据标题和摘要排除明显不相关的论文
+2. **全文阅读**：对相关论文进行深度全文阅读
+3. **维度打分**：按照5个维度逐一打分
+4. **计算总分**：加权计算综合得分
+5. **排序选择**：按总分排序，选出 Top 3
diff --git a/.agents/skills/daily-paper-generator/references/writing-style.md b/.agents/skills/daily-paper-generator/references/writing-style.md
new file mode 100644
index 0000000..856d10f
--- /dev/null
+++ b/.agents/skills/daily-paper-generator/references/writing-style.md
@@ -0,0 +1,134 @@
+# 中文评语写作风格指南
+
+基于 AINet Daily Paper.xlsx 中高质量示例的写作风格总结。
+
+## 评语结构
+
+中文评语约 300 字，遵循以下结构：
+
+```
+1. 背景 (1-2句)     ──> 介绍研究领域的背景和重要性
+2. 挑战 (2-3句)    ──> 指出现有方法面临的关键问题
+3. 贡献 (1-2句)    ──> 概述本工作的核心贡献
+4. 方法 (2-3句)    ──> 描述提出的方法/模型的关键技术
+5. 实验结果 (2-3句) ──> 总结主要实验发现和性能指标
+6. 分析与局限 (1-2句) ──> 分析结果意义，指出局限性
+```
+
+## 常用句式模板
+
+### 开头（背景）
+- `本文针对...问题`
+- `本研究聚焦于...`
+- `本文探讨了...这一基本挑战`
+- `...始终面临诸多挑战`
+
+### 挑战描述
+- `尽管...，但现有方法面临...`
+- `然而，现有方法存在...`
+- `...高度异构，而以往...难以兼顾...`
+- `针对这些问题，本文提出...`
+
+### 方法描述
+- `为此，本文提出...`
+- `该研究提出...`
+- `本文提出...，一种...`
+- `该方法通过...`
+
+### 实验结果
+- `实验结果表明...`
+- `在...数据集上，该方法...`
+- `实验在...上取得了...`
+- `结果表明，...`
+
+### 分析与局限
+- `从而验证了...的可行性`
+- `为...奠定方法学基础`
+- `该工作打破了...`
+- `从而验证了...在实际应用中的可行性`
+
+## 高质量示例分析
+
+### 示例 1：ECHO (Toward Contextual Seq2Seq Paradigms in Large EEG Models)
+
+**结构分析：**
+```
+[背景] 从脑电信号中统一刻画和理解多样化认知任务始终面临诸多挑战
+[挑战] 不同任务形式与标签空间差异显著，不同数据集在电极布局和采集范式上高度异构，
+       而以往以编码器为中心、依赖任务专用预测头的建模方式难以兼顾泛化性与灵活性
+[方法] 为此，本文提出 ECHO，一种以解码器为核心的大规模脑电建模范式，
+       将 EEG 分析重构为统一的序列到序列学习问题
+[技术细节] 该方法将连续脑电片段、任务标识与标签符号共同组织进自回归解码序列中，
+           使模型能够在同一框架下理解任务语境并生成对应预测
+[实验] 在涵盖 12 个公开数据集、6 类脑电任务的统一多任务评测中，
+       ECHO 在整体性能、跨数据集泛化以及零样本场景下均显著优于多种代表性基线方法
+[意义] 为新一代灵活、可扩展的脑机接口系统奠定方法学基础
+```
+
+### 示例 2：EEG-to-Voice Decoding
+
+**结构分析：**
+```
+[背景] 本研究聚焦于从非侵入式 EEG 信号中重建有声语音与想象语音，
+       以辅助存在言语障碍的个体进行交流
+[方法] 其技术框架包括一个被试特异的生成器，以开环方式将预处理后的 EEG 信号映射为 Mel 频谱
+[技术细节] 随后通过预训练的 HiFi-GAN 声码器和 HuBERT ASR 模块生成语音波形并解码文本
+[创新点] 该方法避免了显式的时间对齐过程，采用迁移学习将基于有声语音预训练的生成器
+         适配到想象语音任务中，并使用双损失函数进行训练
+[实验结果] 在有声语音和想象语音两种情形下均实现了稳定的声学重建性能和语言层面的重建性能
+[额外特点] 在语句长度增加的情况下仍能保持文本层面的解码性能
+[意义] 从而验证了 EEG 到语音通信在实际应用中的可行性
+```
+
+## 写作要点
+
+### 1. 学术语言规范
+- 使用正式的学术书面语
+- 避免口语化表达
+- 使用精确的技术术语
+
+### 2. 逻辑连贯
+- 各部分之间有清晰的逻辑关系
+- 使用恰当的连接词（然而、为此、从而、同时）
+- 从问题到解决方案到验证结果
+
+### 3. 信息密度
+- 每句话都承载有效信息
+- 避免冗余表述
+- 突出核心贡献
+
+### 4. 客观评价
+- 准确描述实验结果
+- 指出方法的局限性
+- 不夸大成果
+
+## 英文评语写作
+
+英文评语应与中文评语对应，保持流畅的学术英语风格：
+
+- 使用正式的学术英语
+- 保持与中文评语相同的结构
+- 注意时态一致性（描述论文内容用现在时，描述实验结果用过去时）
+- 使用准确的学术词汇
+
+### 常用句式
+
+**Opening:**
+- `This paper addresses the...`
+- `This study focuses on...`
+- `The paper proposes...`
+
+**Problem:**
+- `While existing...`
+- `However, current methods face...`
+- `Despite progress in...`
+
+**Method:**
+- `To tackle this, the research proposes...`
+- `The authors present...`
+- `This work introduces...`
+
+**Results:**
+- `Experiments demonstrate that...`
+- `The findings show...`
+- `Results indicate that...`
diff --git a/.agents/skills/daily-paper-generator/scripts/arxiv_search.py b/.agents/skills/daily-paper-generator/scripts/arxiv_search.py
new file mode 100755
index 0000000..65dd2d7
--- /dev/null
+++ b/.agents/skills/daily-paper-generator/scripts/arxiv_search.py
@@ -0,0 +1,152 @@
+#!/usr/bin/env python3
+"""
+arXiv 论文搜索脚本
+
+用于搜索 arXiv 上与脑电解码相关的论文。
+
+用法:
+    python arxiv_search.py --query "EEG speech decoding" --max-results 50
+    python arxiv_search.py --keywords EEG speech decoding --months 3
+"""
+
+import argparse
+import feedparser
+import re
+from datetime import datetime, timedelta
+from typing import List, Dict, Optional
+from urllib.parse import quote_plus
+
+
+def search_arxiv(
+    query: str,
+    max_results: int = 50,
+    categories: Optional[List[str]] = None,
+    months: int = 3
+) -> List[Dict]:
+    """
+    搜索 arXiv 论文
+
+    Args:
+        query: 搜索查询字符串
+        max_results: 最大结果数
+        categories: arXiv 分类列表 (如 ['cs.CV', 'cs.LG'])
+        months: 搜索最近几个月的论文
+
+    Returns:
+        论文列表，每个论文包含标题、作者、摘要、链接等信息
+    """
+    # 构建 arXiv API 查询
+    base_url = "http://export.arxiv.org/api/query?"
+
+    # 添加分类过滤
+    if categories:
+        cat_query = " OR ".join([f"cat:{cat}" for cat in categories])
+        search_query = f"search_query=({quote_plus(cat_query)})+AND+all:{quote_plus(query)}"
+    else:
+        search_query = f"search_query=all:{quote_plus(query)}"
+
+    # 其他参数
+    params = f"&start=0&max_results={max_results}&sortBy=submittedDate&sortOrder=descending"
+    url = base_url + search_query + params
+
+    print(f"正在搜索: {url}")
+
+    # 执行查询
+    feed = feedparser.parse(url)
+    papers = []
+
+    # 计算时间截止
+    cutoff_date = datetime.now() - timedelta(days=months * 30)
+
+    for entry in feed.entries:
+        # 解析日期
+        published = datetime(*entry.published_parsed[:6])
+
+        # 时间过滤
+        if published < cutoff_date:
+            continue
+
+        # 解析作者
+        authors = [author.name for author in entry.authors]
+        first_author = authors[0] if authors else "Unknown"
+
+        # 解析 arXiv ID
+        arxiv_id = entry.id.split("/abs/")[-1]
+        arxiv_link = f"https://arxiv.org/abs/{arxiv_id}"
+
+        # 解析摘要（去除多余空白）
+        summary = re.sub(r'\s+', ' ', entry.summary).strip()
+
+        paper = {
+            "title": entry.title,
+            "authors": authors,
+            "first_author": first_author,
+            "summary": summary,
+            "published": published.strftime("%Y-%m-%d"),
+            "arxiv_id": arxiv_id,
+            "arxiv_link": arxiv_link,
+            "pdf_link": f"https://arxiv.org/pdf/{arxiv_id}.pdf",
+            "categories": [tag.term for tag in entry.tags],
+        }
+        papers.append(paper)
+
+    print(f"找到 {len(papers)} 篇相关论文（最近{months}个月）")
+    return papers
+
+
+def print_papers(papers: List[Dict], limit: int = 10):
+    """打印论文列表"""
+    print(f"\n=== 最近 {min(limit, len(papers))} 篇论文 ===\n")
+    for i, paper in enumerate(papers[:limit]):
+        print(f"[{i+1}] {paper['title']}")
+        print(f"    作者: {paper['first_author']} et al.")
+        print(f"    发表: {paper['published']}")
+        print(f"    链接: {paper['arxiv_link']}")
+        print(f"    摘要: {paper['summary'][:150]}...")
+        print()
+
+
+def main():
+    parser = argparse.ArgumentParser(description="搜索 arXiv 论文")
+    parser.add_argument("--query", "-q", type=str, help="搜索查询字符串")
+    parser.add_argument("--keywords", "-k", nargs="+", help="搜索关键词列表")
+    parser.add_argument("--max-results", "-n", type=int, default=50, help="最大结果数")
+    parser.add_argument("--categories", "-c", nargs="+",
+                        default=["cs.CV", "cs.LG", "q-bio.NC"],
+                        help="arXiv 分类")
+    parser.add_argument("--months", "-m", type=int, default=3,
+                        help="搜索最近几个月的论文")
+    parser.add_argument("--output", "-o", type=str, help="输出 JSON 文件路径")
+
+    args = parser.parse_args()
+
+    # 构建查询
+    if args.query:
+        query = args.query
+    elif args.keywords:
+        query = "+".join(args.keywords)
+    else:
+        # 默认查询：脑电语音解码
+        query = "EEG+speech+decoding"
+
+    # 执行搜索
+    papers = search_arxiv(
+        query=query,
+        max_results=args.max_results,
+        categories=args.categories,
+        months=args.months
+    )
+
+    # 打印结果
+    print_papers(papers, limit=10)
+
+    # 输出 JSON
+    if args.output:
+        import json
+        with open(args.output, 'w', encoding='utf-8') as f:
+            json.dump(papers, f, ensure_ascii=False, indent=2)
+        print(f"\n结果已保存到 {args.output}")
+
+
+if __name__ == "__main__":
+    main()