@charzhu/openjaw-agent 0.2.0

package/skills/create-docx.md

@@ -0,0 +1,174 @@
+---
+name: create-docx
+description: "Create, read, and edit Word documents (.docx) with professional formatting."
+whenToUse: "User asks to create a Word document, .docx file"
+---
+
+# Skill: Create Word Document (DOCX)
+
+## Description
+Create, read, and edit Word documents (.docx) with professional formatting, tables, images, headers/footers, and Chinese font support. Use this when user wants a Word document deliverable.
+
+## When to Use
+- User asks to create a Word document, .docx file
+- User says "帮我做个Word文档" or "create a docx"
+- User wants a formatted report/proposal/spec as Word (not PDF)
+- User wants to edit an existing .docx file
+
+## Prerequisites
+- Python with `python-docx` library
+- Install if needed: `pip install python-docx`
+
+## Creating Documents with python-docx
+
+### 1. Basic Setup
+```python
+from docx import Document
+from docx.shared import Inches, Pt, Cm, RGBColor
+from docx.enum.text import WD_ALIGN_PARAGRAPH
+from docx.enum.style import WD_STYLE_TYPE
+
+doc = Document()
+
+# Set default font (Chinese support)
+style = doc.styles['Normal']
+font = style.font
+font.name = 'Microsoft YaHei'
+font.size = Pt(11)
+
+# For East Asian font fallback
+from docx.oxml.ns import qn
+style.element.rPr.rFonts.set(qn('w:eastAsia'), 'Microsoft YaHei')
+```
+
+### 2. Content Elements
+
+#### Headings
+```python
+doc.add_heading('Title', level=0)  # Document title
+doc.add_heading('Section', level=1)  # Major section
+doc.add_heading('Subsection', level=2)  # Minor section
+```
+
+#### Paragraphs
+```python
+p = doc.add_paragraph('Normal paragraph text.')
+p.alignment = WD_ALIGN_PARAGRAPH.JUSTIFY
+
+# Bold, italic, colored text
+p = doc.add_paragraph()
+run = p.add_run('Bold text ')
+run.bold = True
+run = p.add_run('and normal text.')
+```
+
+#### Bullet Lists
+```python
+doc.add_paragraph('First item', style='List Bullet')
+doc.add_paragraph('Second item', style='List Bullet')
+doc.add_paragraph('Numbered item', style='List Number')
+```
+
+#### Tables
+```python
+table = doc.add_table(rows=3, cols=3, style='Light Grid Accent 1')
+# Set header row
+hdr_cells = table.rows[0].cells
+hdr_cells[0].text = 'Column 1'
+hdr_cells[1].text = 'Column 2'
+hdr_cells[2].text = 'Column 3'
+# Add data
+row = table.rows[1].cells
+row[0].text = 'Data 1'
+row[1].text = 'Data 2'
+row[2].text = 'Data 3'
+```
+
+#### Images
+```python
+doc.add_picture('image.png', width=Inches(4))
+```
+
+#### Page Breaks
+```python
+doc.add_page_break()
+```
+
+### 3. Headers and Footers
+```python
+section = doc.sections[0]
+header = section.header
+header_para = header.paragraphs[0]
+header_para.text = "CONFIDENTIAL"
+header_para.alignment = WD_ALIGN_PARAGRAPH.RIGHT
+
+footer = section.footer
+footer_para = footer.paragraphs[0]
+footer_para.text = "Page "
+# Note: auto page numbers require XML manipulation
+```
+
+### 4. Page Setup
+```python
+from docx.shared import Cm
+section = doc.sections[0]
+section.page_height = Cm(29.7)  # A4
+section.page_width = Cm(21.0)
+section.top_margin = Cm(2.54)
+section.bottom_margin = Cm(2.54)
+section.left_margin = Cm(2.54)
+section.right_margin = Cm(2.54)
+```
+
+### 5. Custom Styles
+```python
+# Create a custom style
+custom_style = doc.styles.add_style('CustomQuote', WD_STYLE_TYPE.PARAGRAPH)
+custom_style.font.italic = True
+custom_style.font.color.rgb = RGBColor(0x66, 0x66, 0x66)
+custom_style.font.size = Pt(10)
+custom_style.paragraph_format.left_indent = Cm(1.5)
+```
+
+### 6. Save
+```python
+import os, tempfile
+output_path = os.path.join(tempfile.gettempdir(), 'document.docx')
+doc.save(output_path)
+```
+
+## Reading Existing Documents
+```python
+doc = Document('existing.docx')
+
+# Read all paragraphs
+for para in doc.paragraphs:
+    print(f"[{para.style.name}] {para.text}")
+
+# Read all tables
+for table in doc.tables:
+    for row in table.rows:
+        for cell in row.cells:
+            print(cell.text, end='\t')
+        print()
+```
+
+## Table Styles Available
+Common built-in styles (no need to define):
+- `Table Grid` — Simple borders
+- `Light Grid Accent 1` — Blue header with light grid
+- `Light Shading Accent 1` — Blue shaded header
+- `Medium Shading 1 Accent 1` — Stronger blue styling
+
+## Important Notes
+- **Chinese fonts**: Always set both `font.name` and `w:eastAsia` for proper Chinese rendering
+- **Never use `\n` in paragraphs** — create new Paragraph objects instead
+- **Table widths**: Set column widths explicitly for consistent rendering
+- **Images**: Always specify width to avoid oversized images
+- **Compatibility**: python-docx creates Office-compatible .docx files that work in Word, Google Docs, and LibreOffice
+- **For complex layouts**: Consider using create-pdf-report skill instead (more control over positioning)
+
+## Output & Delivery
+- Save to temp: system temp directory (e.g. `tempfile.gettempdir()`) as `{filename}.docx`
+- Send via email: Use email-with-attachment skill (Outlook REST API)
+  - ContentType: `application/vnd.openxmlformats-officedocument.wordprocessingml.document`

package/skills/create-pdf-report.md

@@ -0,0 +1,93 @@
+---
+name: create-pdf-report
+description: "Generate professional PDF reports with tables, charts, and styled layouts."
+whenToUse: "User asks to create a PDF document or report"
+---
+
+# Skill: Create PDF Report
+
+## Description
+Generate professional PDF reports with Chinese font support, tables, charts, and styled layouts using Python reportlab. Use this skill whenever the user asks to create a PDF, report, analysis document, or any deliverable in PDF format.
+
+## When to Use
+- User asks to create a PDF document or report
+- User asks to generate an analysis, summary, or briefing as a downloadable file
+- User asks to "整理成PDF" or "发给我" with document content
+- Any task requiring formatted document output
+
+## Prerequisites
+- Python with `reportlab` library (confirmed available)
+- Chinese font: `C:/Windows/Fonts/msyh.ttc` (Microsoft YaHei) and `C:/Windows/Fonts/msyhbd.ttc` (Bold)
+
+## Steps
+
+### 1. Register Chinese Fonts
+```python
+from reportlab.pdfbase import pdfmetrics
+from reportlab.pdfbase.ttfonts import TTFont
+
+pdfmetrics.registerFont(TTFont('MSYH', "C:/Windows/Fonts/msyh.ttc", subfontIndex=0))
+pdfmetrics.registerFont(TTFont('MSYHB', "C:/Windows/Fonts/msyhbd.ttc", subfontIndex=0))
+```
+
+### 2. Use ParagraphStyle Helper
+Define a helper function to avoid verbose style definitions:
+```python
+from reportlab.lib.styles import ParagraphStyle
+from reportlab.lib.enums import TA_LEFT, TA_CENTER, TA_JUSTIFY
+
+def S(name, **kw):
+    defaults = dict(fontName='MSYH', fontSize=10.5, leading=18, textColor=DARK)
+    defaults.update(kw)
+    return ParagraphStyle(name, **defaults)
+```
+
+### 3. Standard Color Palette
+```python
+from reportlab.lib.colors import HexColor, white
+
+PRIMARY = HexColor('#1a73e8')   # Blue
+DARK = HexColor('#1a1a2e')      # Dark text
+ACCENT = HexColor('#e94560')    # Red accent
+GREEN = HexColor('#27ae60')     # Green
+ORANGE = HexColor('#f39c12')    # Orange
+GRAY = HexColor('#666666')      # Gray text
+LGRAY = HexColor('#e0e0e0')     # Light gray borders
+LIGHT_BLUE = HexColor('#e8f0fe')
+LIGHT_GREEN = HexColor('#e8f5e9')
+LIGHT_RED = HexColor('#fce4ec')
+```
+
+### 4. Document Structure
+```python
+from reportlab.lib.pagesizes import A4
+from reportlab.lib.units import mm, cm
+from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, Table, TableStyle, PageBreak, HRFlowable
+
+doc = SimpleDocTemplate(output_path, pagesize=A4, 
+    rightMargin=2*cm, leftMargin=2*cm, topMargin=2*cm, bottomMargin=2*cm)
+story = []
+```
+
+### 5. Table Best Practices
+- Always use Paragraph objects inside table cells (not plain strings) for Chinese text
+- Set VALIGN to 'MIDDLE' or 'TOP'
+- Use alternating row backgrounds for readability
+- Header rows: colored background + white text
+
+### 6. Output Path
+- Save to: `~/Desktop/{filename}.pdf` or system temp directory as `{filename}.pdf`
+- Include full file path in response message so bridge can auto-send
+
+### 7. Important Notes
+- **Encoding**: Use `sys.stdout.reconfigure(encoding='utf-8')` if printing Chinese to stdout
+- **Emoji**: Avoid emoji in print statements (cp1252 encoding error). Use emoji only in PDF content via Paragraph objects
+- **String escaping**: Be careful with nested quotes in Chinese strings — use unicode escapes or triple-quoted strings
+- **NEVER use Unicode subscript/superscript characters** in reportlab — use `<sub>` and `<super>` tags instead
+
+## Verification
+After generating, verify with:
+```python
+print(f"PDF saved to: {output_path}")
+print(f"File size: {os.path.getsize(output_path) / 1024:.1f} KB")
+```

package/skills/create-pptx.md

@@ -0,0 +1,109 @@
+---
+name: create-pptx
+description: "Generate professional PowerPoint presentations with bold design and structured layouts."
+whenToUse: "User asks to create a presentation, deck, or slides"
+---
+
+# Skill: Create PowerPoint Presentation
+
+## Description
+Generate professional PowerPoint presentations (.pptx) with bold design, Chinese support, and structured layouts. Use this skill whenever the user asks for a deck, slides, presentation, or PPT.
+
+## When to Use
+- User asks to create a presentation, deck, or slides
+- User asks to make a PPT/PPTX
+- User says "做个PPT" or "帮我做slides"
+- Any deliverable that should be in presentation format
+
+## Prerequisites
+- Python with `python-pptx` library
+- Install if needed: `pip install python-pptx`
+
+## Design Principles (from Anthropic Skills)
+- **Don't create boring slides** — plain bullets on white background won't impress
+- **Every slide needs a visual element** — image, chart, icon, or shape
+- **Pick a bold color palette** specific to the topic
+- **Dominance over equality** — one color dominates (60-70%), with 1-2 supporting tones
+- **Dark/light contrast** — dark backgrounds for title + conclusion, light for content
+
+## Color Palettes
+
+| Theme | Primary | Secondary | Accent |
+|-------|---------|-----------|--------|
+| Midnight Executive | `1E2761` | `CADCFC` | `FFFFFF` |
+| Ocean Gradient | `065A82` | `1C7293` | `21295C` |
+| Coral Energy | `F96167` | `F9E795` | `2F3C7E` |
+| Charcoal Minimal | `36454F` | `F2F2F2` | `212121` |
+| Teal Trust | `028090` | `00A896` | `02C39A` |
+
+## Typography
+| Element | Size |
+|---------|------|
+| Slide title | 36-44pt bold |
+| Section header | 20-24pt bold |
+| Body text | 14-16pt |
+| Captions | 10-12pt muted |
+
+## Steps
+
+### 1. Basic Setup
+```python
+from pptx import Presentation
+from pptx.util import Inches, Pt, Emu
+from pptx.dml.color import RGBColor
+from pptx.enum.text import PP_ALIGN, MSO_ANCHOR
+
+prs = Presentation()
+prs.slide_width = Inches(13.333)  # 16:9 widescreen
+prs.slide_height = Inches(7.5)
+```
+
+### 2. Chinese Font Support
+```python
+from pptx.util import Pt
+
+def set_font(run, name='Microsoft YaHei', size=14, bold=False, color=None):
+    run.font.name = name
+    run.font.size = Pt(size)
+    run.font.bold = bold
+    if color:
+        run.font.color.rgb = RGBColor(*color)
+    # East Asian font fallback
+    run.font._element.attrib['{http://schemas.openxmlformats.org/drawingml/2006/main}altLang'] = 'zh-CN'
+```
+
+### 3. Add Shaped Backgrounds
+```python
+from pptx.util import Inches
+from pptx.enum.shapes import MSO_SHAPE
+
+def add_bg_rect(slide, left, top, width, height, color):
+    shape = slide.shapes.add_shape(
+        MSO_SHAPE.RECTANGLE, left, top, width, height
+    )
+    shape.fill.solid()
+    shape.fill.fore_color.rgb = RGBColor(*color)
+    shape.line.fill.background()  # No border
+    return shape
+```
+
+### 4. Slide Layouts
+- **Title slide**: Full dark background, centered title + subtitle
+- **Section divider**: Half-colored, half-white with large section number
+- **Content slide**: Header bar + content area with bullet points or grid
+- **Two-column**: Text left, visual right (or vice versa)
+- **Data slide**: Large stat callouts (60-72pt numbers) with small labels
+
+### 5. Save & Output
+```python
+import os
+output_path = os.path.join(os.path.expanduser('~'), 'Desktop', 'presentation.pptx')
+prs.save(output_path)
+```
+
+## Important Notes
+- Use `Inches()` or `Emu()` for positioning — never raw numbers
+- 0.5" minimum margins
+- Leave breathing room — don't fill every inch
+- **Avoid text-only slides** — always add at least one visual element
+- For complex presentations, consider creating a master template first

package/skills/daily-briefing.md

@@ -0,0 +1,65 @@
+---
+name: daily-briefing
+description: "Gives a quick morning overview of emails, calendar, and Teams activity."
+whenToUse: "User asks for a morning briefing or daily summary of emails, calendar, and Teams"
+---
+
+# Skill: Daily Briefing
+
+## Description
+Gives a quick morning overview of emails, calendar, and Teams activity.
+
+## Channel Fallback Strategy
+For both Outlook and Teams, try channels in this order:
+1. **Graph** (preferred — fast, structured API access)
+2. **Desktop** (fallback — COM/UIA automation, works offline)
+3. **Web** (last resort — browser automation, requires login)
+
+If a channel fails (e.g., token error), immediately switch to the next one and retry.
+Do NOT stop or report failure until all channels have been tried.
+
+## Steps
+
+1. **Email Summary**
+   - Try: `outlook_switch_channel("graph")` → `outlook_go_to_inbox` → `outlook_read_content`
+   - If Graph fails: `outlook_switch_channel("desktop")` → retry
+   - If Desktop fails: `outlook_switch_channel("web")` → navigate to outlook.office.com/mail/inbox → retry
+   - Read the latest 10 emails
+   - Summarize: who sent what, flag anything urgent (action required, approvals, incidents)
+
+2. **Calendar Overview**
+   - Try: `outlook_calendar` → `outlook_read_content` (or `browser_extract` on web)
+   - Same fallback order: Graph → Desktop → Web
+   - List today's meetings/events
+   - Highlight any conflicts or upcoming meetings in the next 2 hours
+
+3. **Teams Check**
+   - Try: `teams_switch_channel("graph")` → `teams_search_navigate("PM Team")` → `teams_read_messages`
+   - If Graph fails: `teams_switch_channel("desktop")` → retry
+   - If Desktop fails: `teams_switch_channel("web")` → retry
+   - Check recent messages in key chats (PM team, etc.)
+   - Summarize any unread or @mentions
+
+4. **Output**
+   - Present a concise briefing in this format:
+
+```
+## 📧 Email (X unread)
+- [Urgent] Sender — Subject
+- Sender — Subject
+- ...
+
+## 📅 Calendar (X events today)
+- 10:00 AM — Meeting Name (with Who)
+- 2:00 PM — Meeting Name (with Who)
+- ...
+
+## 💬 Teams Activity
+- Chat Name: Person said "..."
+- Chat Name: X new messages
+- ...
+
+## ⚡ Action Items
+- Reply to X about Y
+- Prepare for Z meeting at 2 PM
+```

package/skills/data-analysis.md

@@ -0,0 +1,127 @@
+---
+name: data-analysis
+description: "Analyze data from various sources and generate visualizations with Chinese label support."
+whenToUse: "User asks to analyze data, metrics, or trends"
+---
+
+# Skill: Data Analysis & Visualization
+
+## Description
+Analyze data from various sources (CSV, Excel, JSON, APIs) and generate visualizations (charts, graphs, tables) with Chinese label support. Output as images, PDFs, or Excel files.
+
+## When to Use
+- User asks to analyze data, metrics, or trends
+- User asks to create charts, graphs, or dashboards
+- User wants to compare numbers, compute statistics
+- User says "帮我分析" or "做个图表"
+
+## Prerequisites
+- Python: pandas, matplotlib, numpy (confirmed available)
+- Optional: openpyxl (for Excel), seaborn (for advanced plots)
+
+## Steps
+
+### 1. Chinese Font Setup for Matplotlib
+```python
+import matplotlib.pyplot as plt
+import matplotlib
+
+# Set Chinese font
+matplotlib.rcParams['font.sans-serif'] = ['Microsoft YaHei']
+matplotlib.rcParams['axes.unicode_minus'] = False  # Fix minus sign
+```
+
+### 2. Read Data
+```python
+import pandas as pd
+
+# CSV
+df = pd.read_csv('data.csv')
+
+# Excel
+df = pd.read_excel('data.xlsx', sheet_name='Sheet1')
+
+# JSON
+df = pd.read_json('data.json')
+
+# From clipboard
+df = pd.read_clipboard()
+```
+
+### 3. Quick Analysis
+```python
+# Overview
+print(df.shape)           # (rows, cols)
+print(df.describe())      # Statistics
+print(df.info())          # Types & nulls
+print(df.head(10))        # Preview
+
+# Group by
+summary = df.groupby('category').agg({'value': ['mean', 'sum', 'count']})
+
+# Pivot table
+pivot = pd.pivot_table(df, values='value', index='date', columns='category', aggfunc='mean')
+```
+
+### 4. Visualization Templates
+
+#### Bar Chart
+```python
+fig, ax = plt.subplots(figsize=(10, 6))
+df.plot(kind='bar', x='category', y='value', ax=ax, color='#1a73e8')
+ax.set_title('标题', fontsize=16, fontweight='bold')
+ax.set_xlabel('类别')
+ax.set_ylabel('数值')
+plt.tight_layout()
+plt.savefig('chart.png', dpi=150, bbox_inches='tight')
+```
+
+#### Line Chart (Time Series)
+```python
+fig, ax = plt.subplots(figsize=(12, 6))
+for col in columns:
+    ax.plot(df['date'], df[col], marker='o', label=col)
+ax.legend()
+ax.set_title('趋势分析')
+ax.grid(True, alpha=0.3)
+plt.tight_layout()
+plt.savefig('trend.png', dpi=150)
+```
+
+#### Pie Chart
+```python
+fig, ax = plt.subplots(figsize=(8, 8))
+ax.pie(values, labels=labels, autopct='%1.1f%%', colors=palette)
+ax.set_title('占比分析')
+plt.savefig('pie.png', dpi=150)
+```
+
+#### Heatmap
+```python
+import seaborn as sns
+fig, ax = plt.subplots(figsize=(10, 8))
+sns.heatmap(correlation_matrix, annot=True, cmap='RdYlBu_r', center=0, ax=ax)
+plt.tight_layout()
+plt.savefig('heatmap.png', dpi=150)
+```
+
+### 5. Output Options
+- **PNG image**: `plt.savefig('chart.png', dpi=150)` — good for quick sharing
+- **PDF report**: Combine with create-pdf-report skill
+- **Excel**: `df.to_excel('output.xlsx', index=False)`
+- **HTML table**: `df.to_html('table.html')`
+
+### 6. Color Palettes
+```python
+# Microsoft-style
+MS_COLORS = ['#1a73e8', '#e94560', '#27ae60', '#f39c12', '#9b59b6', '#1abc9c']
+
+# Muted professional
+PRO_COLORS = ['#264653', '#2a9d8f', '#e9c46a', '#f4a261', '#e76f51']
+```
+
+## Important Notes
+- Always use `sys.stdout.reconfigure(encoding='utf-8')` for Chinese output
+- Save charts to Desktop or Temp for bridge auto-send
+- Close figures after saving: `plt.close(fig)` to avoid memory leaks
+- For large datasets, sample first: `df.sample(1000)` for visualization

package/skills/deep-research.md

@@ -0,0 +1,101 @@
+---
+name: deep-research
+description: "Conduct in-depth, multi-step research by decomposing queries and synthesizing reports."
+whenToUse: "User asks for thorough research on a topic, technology, or competitor"
+---
+
+# Skill: Deep Research
+
+## Description
+Conduct in-depth, multi-step research on a given topic by decomposing queries, gathering from diverse sources, cross-referencing, and synthesizing into a comprehensive report. Goes beyond simple web search.
+
+## When to Use
+- User asks for thorough research on a topic, technology, or competitor
+- User says "深入调研一下" or "do a deep dive on"
+- User needs more than a quick answer — needs analysis with sources
+- Replaces and enhances the basic web-research skill for complex queries
+
+## Workflow
+
+### 1. Decompose the Query
+Break the main question into 5-8 sub-questions:
+```
+Main: "How is AI changing the PM role?"
+Sub-questions:
+  1. What PM tasks can AI automate today?
+  2. What new skills do PMs need in AI era?
+  3. Which companies are restructuring PM teams?
+  4. What do industry reports say about PM job outlook?
+  5. What tools are PMs using for AI-assisted work?
+  6. What are the risks of over-reliance on AI for PMs?
+```
+
+### 2. Multi-Source Research
+For each sub-question:
+```
+web_search(sub_question)                    # Get initial results
+web_search(sub_question + "2026")           # Recent info
+web_search(sub_question + "research paper") # Academic sources
+web_fetch(url, max_length=8000)             # Deep dive on key sources
+```
+
+Source diversity checklist:
+- [ ] Industry reports (McKinsey, Gartner, Forrester)
+- [ ] Tech company blogs (Google AI, Microsoft, Anthropic)
+- [ ] News articles (TechCrunch, The Verge, Wired)
+- [ ] Academic/research papers
+- [ ] Community discussions (HackerNews, Reddit, Twitter/X)
+- [ ] Product documentation / changelogs
+
+### 3. Cross-Reference & Validate
+- Check claims against multiple sources
+- Note conflicting viewpoints
+- Distinguish facts, opinions, and predictions
+- Check source dates — flag anything > 6 months old
+- Note potential biases (vendor reports favor their products)
+
+### 4. Synthesize Report
+```markdown
+# [Topic] Deep Research Report
+**Date:** [Date] | **Researcher:** Jon
+
+## Executive Summary
+[3-5 sentences capturing the key findings]
+
+## Key Findings
+
+### Finding 1: [Title]
+**Evidence:** [What the data/sources say]
+**Implication:** [What this means for us]
+**Confidence:** High/Medium/Low
+**Sources:** [1], [2]
+
+### Finding 2: [Title]
+...
+
+## Conflicting Viewpoints
+- [View A] vs [View B] — [Analysis of which is more credible and why]
+
+## Gaps in Available Information
+- [What we couldn't find or verify]
+
+## Recommendations
+1. [Actionable recommendation based on findings]
+2. [Actionable recommendation based on findings]
+
+## Sources
+1. [Title](URL) — [Organization], [Date]
+2. [Title](URL) — [Organization], [Date]
+```
+
+### 5. Quality Checks
+- Every claim has a source
+- Recent data prioritized over old
+- Balanced viewpoints represented
+- Clear separation of fact vs opinion
+- Actionable recommendations included
+
+### 6. Output Options
+- Direct message response (for quick summaries)
+- PDF report (use create-pdf-report skill)
+- Email with attachment (use email-with-attachment skill)