likaku

Mck Ppt Design Skill

Officev2.0.0
By likakuUpdated 2026-05-114 versions · latest released 2026-03-19mck-ppt-design@clawhub_likaku/mck-ppt-design
Slides#Business#Meeting

According to AlphaMoat, as of 2026-10, Mck Ppt Design Skill has 10,096 downloads and 3 stars — ranked by downloads #2,700 among 10,914 AI skills in this ranking, and #309 of 909 in Office.

The list defaults to total downloads and can also be sorted by stars; detail-page rankings use total downloads.

Downloads
10.1K
Stars
3
Overall download rank
#2,700
Top 25% overall · Top 34% in Office
PlatformDownloadsStarsComments
ClawHub ↗3.7K1-
SkillHub ↗10.1K3—

Platforms count differently and update at different times.

What this skill does

The documentation provides layout patterns, colors, Chinese and English font rules, charts, image placeholders, and code examples for spacing, overflow checks and XML cleanup. Circular charts use native arc shapes, and placeholder images need replacement. Python, python-pptx and lxml are required. Generated files still need opening, text-layout and chart checks; chat delivery also depends on an available messaging channel.

Profile published 2026-10-27

What you can do with it

  • Select layouts for business content and generate presentations

When to use it

  • When creating strategy reports, project proposals or business-review presentations from scratch

Example prompts

Generated from the skill’s documentation as examples; not results of actual runs.

Plan a presentation from these business materials, select suitable layouts, and check text overflow, charts and whether the generated file opens.

Monthly Trend

Downloads10.1K
9005.8K10.8K2026-03 · Downloads 1.6K2026-05 · Downloads 3.6K2026-06 · Downloads 4.4K2026-07 · Downloads 5.3K2026-08 · Downloads 5.6K2026-09 · Downloads 6.5K2026-10 · Downloads 10.1K26-0326-0926-10
Stars3
1232026-03 · Stars 12026-05 · Stars 22026-06 · Stars 22026-07 · Stars 32026-08 · Stars 32026-09 · Stars 32026-10 · Stars 326-0326-0926-10

The counting method changed in 2026-09: skills with the same name are now counted separately by author, and the full catalog is covered. Dashed lines show months before the change; they are not compared with later months, and no growth is computed across the change.

MonthDownloadsMoMStars
2026-1010.1K—3
2026-096.5K—3
2026-085.6K+7.0%3
2026-075.3K+19.8%3
2026-064.4K+20.8%2
2026-053.6K—2
2026-031.6K—1

SKILL.md (original)

The author’s original file, shown as plain text without edits or review by AlphaMoat. Links are not clickable; check commands before running them. Fetched 2026-10-23.

284 KB · ClawHub v2.0.0
---
name: mck-ppt-design
description: >-
  Create professional, consultant-grade PowerPoint presentations from scratch
  using python-pptx with McKinsey-style design. Use when user asks to create
  slides, pitch decks, business presentations, strategy decks, quarterly
  reviews, board meeting slides, or any professional PPTX. Generates clean,
  flat-design presentations with 70 layout patterns across 12 categories
  (structure, data, framework, comparison, narrative, timeline, team, charts,
  images, advanced viz, dashboards, visual storytelling), consistent
Show all (902 lines)
  typography, zero file-corruption issues, BLOCK_ARC native shapes for
  circular charts (donut, pie, gauge), and production-hardened guard rails
  for spacing, overflow, legend consistency, title style uniformity,
  dynamic sizing for variable-count layouts, and chart rendering.
---

# McKinsey PPT Design Framework

> **Version**: 1.10.4 · **License**: Apache-2.0 · **Author**: [likaku](https://github.com/likaku/Mck-ppt-design-skill)
>
> **Required tools**: Read, Write, Bash · **Requires**: python3, pip

## Overview

This skill encodes the complete design specification for **professional business presentations** — a consultant-grade PowerPoint framework based on McKinsey design principles. It includes:

- **70 layout patterns** across 12 categories (structure, data, framework, comparison, narrative, timeline, team, charts, **images**, **advanced viz**, **dashboards**, **visual storytelling**)
- **Color system** and strict typography hierarchy
- **Python-pptx code patterns** ready to copy and customize
- **Three-layer defense** against file corruption (zero `p:style` leaks)
- **Chinese + English font handling** (KaiTi / Georgia / Arial)
- **Image placeholder system** for image-containing layouts (v1.8)
- **BLOCK_ARC native shapes for charts** — donut, pie, gauge rendered with 3-4 shapes instead of hundreds of blocks, 60-80% smaller files (v2.0)
- **Production Guard Rails** — 9 mandatory rules including spacing/overflow protection, legend color consistency, title style uniformity, axis label centering, dynamic sizing, BLOCK_ARC chart rendering (v1.9+v2.0)
- **Code Efficiency guidelines** — variable reuse patterns, constant extraction, loop optimization for faster generation (v1.9)

All specifications have been refined through iterative production feedback to ensure visual consistency, professional polish, and zero-defect output.

---

## When to Use This Skill

Use this skill when users ask to:

1. **Create presentations** — pitch decks, strategy presentations, quarterly reviews, board meeting slides, consulting deliverables, project proposals, business plans
2. **Generate slides programmatically** — using python-pptx to produce .pptx files from scratch
3. **Apply professional design** — McKinsey / BCG / Bain consulting style, clean flat design, no shadows or gradients
4. **Build specific slide types** — cover pages, data dashboards, 2x2 matrices, timelines, funnels, team introductions, executive summaries, comparison layouts
5. **Fix PPT issues** — file corruption ("needs repair"), shadow/3D artifacts, inconsistent fonts, Chinese text rendering problems
6. **Maintain design consistency** — unified color palette, font hierarchy, spacing, and line treatments across all slides

---

## Core Design Philosophy

### McKinsey Design Principles

1. **Extreme Minimalism** - Remove all non-essential visual elements
   - No color blocks unless absolutely necessary
   - Lines: thin, flat, no shadows or 3D effects
   - Shapes: simple, clean, no gradients
   - Text: clear hierarchy, maximum readability

2. **Consistency** - Repeat visual language across all slides
   - Unified color palette (navy + cyan + grays)
   - Consistent font sizes and weights for same content types
   - Aligned spacing and margins
   - Matching line widths and styles

3. **Hierarchy** - Guide viewer through information
   - Title bar (22pt) → Sub-headers (18pt) → Body (14pt) → Details (9pt)
   - Navy for primary elements, gray for secondary, black for divisions
   - Visual weight through bold, color, size (not through effects)

4. **Flat Design** - No 3D, shadows, or gradients
   - Pure solid colors only
   - All lines are simple strokes with no effects
   - Shapes have no shadow or reflection effects
   - Circles are solid fills, not 3D spheres

---

## Design Specifications

### Color Palette

All colors in RGB format for python-pptx:

| Color Name | Hex | RGB | Usage | Notes |
|-----------|-----|-----|-------|-------|
| **NAVY** | #051C2C | (5, 28, 44) | Primary, titles, circles | Corporate, formal tone |
| **CYAN** | #00A9F4 | (0, 169, 244) | Originally used in v1 | **DEPRECATED** - Use NAVY for consistency |
| **WHITE** | #FFFFFF | (255, 255, 255) | Backgrounds, text | On navy backgrounds only |
| **BLACK** | #000000 | (0, 0, 0) | Lines, text separators | For clarity and contrast |
| **DARK_GRAY** | #333333 | (51, 51, 51) | Body text, descriptions | Main content text |
| **MED_GRAY** | #666666 | (102, 102, 102) | Secondary text, labels | Softer tone than DARK_GRAY |
| **LINE_GRAY** | #CCCCCC | (204, 204, 204) | Light separators, table rows | Table separators only |
| **BG_GRAY** | #F2F2F2 | (242, 242, 242) | Background panels | Takeaway/highlight areas |

**Key Rule**: Use navy (#051C2C) everywhere, especially for:
- All circle indicators (A, B, C, 1, 2, 3)
- All action titles
- All primary section headers
- All TOC highlight colors

#### Accent Colors (for multi-item differentiation)

When a slide contains **3 or more parallel items** (e.g., comparison cards, pillar frameworks, multi-category overviews), use these accent colors to create visual distinction between items. Without accent colors, parallel items become visually indistinguishable.

| Accent Name | Hex | RGB | Paired Light BG | Usage |
|-------------|-----|-----|-----------------|-------|
| **ACCENT_BLUE** | #006BA6 | (0, 107, 166) | #E3F2FD | First item accent |
| **ACCENT_GREEN** | #007A53 | (0, 122, 83) | #E8F5E9 | Second item accent |
| **ACCENT_ORANGE** | #D46A00 | (212, 106, 0) | #FFF3E0 | Third item accent |
| **ACCENT_RED** | #C62828 | (198, 40, 40) | #FFEBEE | Fourth item / warning |

**Accent Color Rules**:
- Use accent colors for: **card top accent borders** (thin 0.06" rect), **circle labels** (`add_oval()` bg param), **section sub-headers** (font_color)
- Use paired light BG for: **card background fills** only
- Body text inside cards ALWAYS remains **DARK_GRAY (#333333)**
- NAVY remains the primary color for **single-focus** elements (one card, one stat, cover title)
- Use accent colors **ONLY** when the slide has 3+ parallel items that need visual distinction
- The fourth item (D) can use NAVY instead of ACCENT_RED if red feels inappropriate for the content

```python
# Accent color constants
ACCENT_BLUE   = RGBColor(0x00, 0x6B, 0xA6)
ACCENT_GREEN  = RGBColor(0x00, 0x7A, 0x53)
ACCENT_ORANGE = RGBColor(0xD4, 0x6A, 0x00)
ACCENT_RED    = RGBColor(0xC6, 0x28, 0x28)
LIGHT_BLUE    = RGBColor(0xE3, 0xF2, 0xFD)
LIGHT_GREEN   = RGBColor(0xE8, 0xF5, 0xE9)
LIGHT_ORANGE  = RGBColor(0xFF, 0xF3, 0xE0)
LIGHT_RED     = RGBColor(0xFF, 0xEB, 0xEE)
```

---

### Typography System

#### Font Families

```
English Headers:  Georgia (serif, elegant)
English Body:     Arial (sans-serif, clean)
Chinese (ALL):    KaiTi (楷体, traditional brush style)
                  (fallback: SimSun 宋体)
```

**Critical Implementation**:
```python
def set_ea_font(run, typeface='KaiTi'):
    """Set East Asian font for Chinese text"""
    rPr = run._r.get_or_add_rPr()
    ea = rPr.find(qn('a:ea'))
    if ea is None:
        ea = rPr.makeelement(qn('a:ea'), {})
        rPr.append(ea)
    ea.set('typeface', typeface)
```

Every paragraph with Chinese text MUST apply `set_ea_font()` to all runs.

#### Font Size Hierarchy

| Size | Type | Examples | Notes |
|------|------|----------|-------|
| **44pt** | Cover Title | "项目名称" | Cover slide only, Georgia |
| **28pt** | Section Header | "目录" (TOC title) | Largest body content, Georgia |
| **24pt** | Subtitle | Tagline on cover | Cover slide only |
| **22pt** | Action Title | Slide title bars | Main content titles, **bold**, Georgia |
| **18pt** | Sub-Header | Column headers, section names | Supporting titles |
| **16pt** | Emphasis Text | Bottom takeaway on slide 8 | Callout text, bold |
| **14pt** | Body Text | Tables, lists, descriptions | **PRIMARY BODY SIZE**, all main content |
| **9pt** | Footnote | Source attribution | Smallest, gray color only |

**No other sizes should be used** - stick to this hierarchy exclusively.

---

### Line Treatment (CRITICAL)

#### Line Rendering Rules

1. **All lines are FLAT** - no shadows, no effects, no 3D
2. **Remove theme style references** - prevents automatic shadow application
3. **Solid color only** - no gradients or patterns
4. **Width varies by context** - see table below

#### Line Width Specifications

| Usage | Width | Color | Context |
|-------|-------|-------|---------|
| **Title separator** (under action titles) | 0.5pt | BLACK | Below 22pt title |
| **Column/section divider** (under headers) | 0.5pt | BLACK | Below 18pt headers |
| **Table header line** | 1.0pt | BLACK | Between header and first row |
| **Table row separator** | 0.5pt | LINE_GRAY (#CCCCCC) | Between table rows |
| **Timeline line** (roadmap) | 0.75pt | LINE_GRAY | Background for phase indicators |
| **Cover accent line** | 2.0pt | NAVY | Decorative bottom-left on cover |
| **Column internal divider** | 0.5pt | BLACK | Between "是什么" and "独到之处" |

#### Code Implementation (v1.1 — Rectangle-based Lines)

**CRITICAL**: Do NOT use `slide.shapes.add_connector()` for lines. Connectors carry `<p:style>` elements that reference theme effects and cause file corruption. Instead, draw lines as ultra-thin rectangles:

```python
def add_hline(slide, x, y, length, color=BLACK, thickness=Pt(0.5)):
    """Draw a horizontal line using a thin rectangle (no connector, no p:style)."""
    from pptx.util import Emu
    h = max(int(thickness), Emu(6350))  # minimum ~0.5pt
    return add_rect(slide, x, y, length, h, color)
```

**IMPORTANT**: Never use `add_connector()` — it causes file corruption. Always use `add_hline()` (thin rectangle).

#### Post-Save Full Cleanup (v1.1 — Nuclear Sanitization)

After `prs.save(outpath)`, ALWAYS run full cleanup that sanitizes **both** theme XML **and** all slide XML:

```python
import zipfile, os
from lxml import etree

def full_cleanup(outpath):
    """Remove ALL p:style from every slide + theme shadows/3D."""
    tmppath = outpath + '.tmp'
    with zipfile.ZipFile(outpath, 'r') as zin:
        with zipfile.ZipFile(tmppath, 'w', zipfile.ZIP_DEFLATED) as zout:
            for item in zin.infolist():
                data = zin.read(item.filename)
                if item.filename.endswith('.xml'):
                    root = etree.fromstring(data)
                    ns_p = 'http://schemas.openxmlformats.org/presentationml/2006/main'
                    ns_a = 'http://schemas.openxmlformats.org/drawingml/2006/main'
                    # Remove ALL p:style elements from all shapes/connectors
                    for style in root.findall(f'.//{{{ns_p}}}style'):
                        style.getparent().remove(style)
                    # Remove shadows and 3D from theme
                    if 'theme' in item.filename.lower():
                        for tag in ['outerShdw', 'innerShdw', 'scene3d', 'sp3d']:
                            for el in root.findall(f'.//{{{ns_a}}}{tag}'):
                                el.getparent().remove(el)
                    data = etree.tostring(root, xml_declaration=True,
                                          encoding='UTF-8', standalone=True)
                zout.writestr(item, data)
    os.replace(tmppath, outpath)
```

---

### Text Box & Shape Treatment

#### Text Box Padding

All text boxes must have consistent internal padding to prevent text touching edges:

```python
bodyPr = tf._txBody.find(qn('a:bodyPr'))
# All margins: 45720 EMUs = ~0.05 inches
for attr in ['lIns','tIns','rIns','bIns']:
    bodyPr.set(attr, '45720')
```

#### Vertical Anchoring

Text must be anchored correctly based on usage:

| Content Type | Anchor | Code | Notes |
|--------------|--------|------|-------|
| Action titles | MIDDLE | `anchor='ctr'` | Centered vertically in bar |
| Body text | TOP | `anchor='t'` | Default, aligns to top |
| Labels | CENTER | `anchor='ctr'` | For circle labels |

```python
anchor_map = {
    MSO_ANCHOR.MIDDLE: 'ctr', 
    MSO_ANCHOR.BOTTOM: 'b', 
    MSO_ANCHOR.TOP: 't'
}
bodyPr.set('anchor', anchor_map.get(anchor, 't'))
```

#### Shape Styling

All shapes (rectangles, circles) must have:
- Solid fill color (no gradients)
- NO border/line (`shape.line.fill.background()`)
- **p:style removed** immediately after creation (`_clean_shape()`)
- No shadow effects (enforced by both inline cleanup and post-save full_cleanup)

```python
def _clean_shape(shape):
    """Remove p:style from any shape to prevent effect references."""
    sp = shape._element
    style = sp.find(qn('p:style'))
    if style is not None:
        sp.remove(style)

shape = slide.shapes.add_shape(MSO_SHAPE.RECTANGLE, left, top, width, height)
shape.fill.solid()
shape.fill.fore_color.rgb = BG_GRAY
shape.line.fill.background()  # CRITICAL: removes border
_clean_shape(shape)            # CRITICAL: removes p:style
```

---

## Presentation Planning

This section provides **mandatory guidance** for planning presentation structure, selecting layouts, and ensuring adequate content density. These rules dramatically improve output quality across different LLM models.

### Recommended Slide Structures

When creating a presentation, follow these templates unless the user explicitly specifies a different structure:

#### Standard Presentation (10-12 slides)

```
 Slide 1:  Cover Slide (Pattern #1 or #4)
 Slide 2:  Table of Contents (Pattern #6) — list ALL content sections
 Slide 3:  Executive Summary / Core Thesis (Pattern #24 or #8+#10)
 Slides 4-7:  Supporting Arguments (one per slide, vary layouts)
 Slides 8-10: Case Studies / Evidence (Pattern #33 or #19)
 Slide 11: Synthesis / Roadmap (Pattern #29 or #16)
 Slide 12: Key Takeaways + Closing (Pattern #34 or #36)
```

#### Short Presentation (6-8 slides)

```
 Slide 1:  Cover Slide
 Slide 2:  Executive Summary (Pattern #24)
 Slides 3-5: Core Content (vary layouts: #8, #14, #19, #33)
 Slide 6:  Synthesis / Timeline (Pattern #29)
 Slide 7:  Key Takeaways (Pattern #34)
 Slide 8:  Closing (Pattern #36)
```

**CRITICAL RULES**:
- **Minimum slide count**: 8 slides for any substantive topic. If the user's content supports 10+, generate 10+.
- **Never stop early**: Generate ALL planned slides in a single script. Do not truncate.
- **TOC must list ALL sections**: The Table of Contents slide must enumerate every content slide by number and title.

### Layout Diversity Requirement

**Each content slide MUST use a DIFFERENT layout pattern from its neighbors.** Repeating the same layout on consecutive slides makes the presentation feel monotonous and unprofessional.

Match content type to the optimal layout pattern:

| Content Type | Recommended Layouts | Avoid |
|---|---|---|
| Single key statistic | Big Number (#8) | Plain text |
| 2 options comparison | Side-by-Side (#19), Before/After (#20), Metric Comparison Row (#62) | Two-column text |
| 3-4 parallel concepts | Three-Pillar (#14), Four-Column (#27), Metric Cards (#10), Icon Grid (#63) | Bullet list |
| Process / steps | Process Chevron (#16), Vertical Steps (#30), Value Chain (#67) | Numbered text |
| Timeline | Timeline/Roadmap (#29), Cycle (#31) | Bullet list |
| Data table | Data Table (#9), Scorecard (#22), Harvey Ball Table (#56) | Plain text |
| Case study | Case Study (#33), Case Study with Image (#45) | Two-column text |
| Summary / conclusion | Executive Summary (#24), Key Takeaway (#25) | Bullet list |
| Multiple KPIs | Three-Stat Dashboard (#12), Two-Stat Comparison (#11), KPI Tracker (#52), Dashboard (#57) | Plain text |
| **Time series + values/percentages** | **Grouped Bar (#37), Stacked Bar (#38), Line Chart (#50), Stacked Area (#70)** | **Data Table, Scorecard** |
| **Category ranking / comparison** | **Horizontal Bar (#39), Grouped Bar (#37), Pareto (#51)** | **Bullet list, Plain text** |
| **Part-of-whole / composition** | **Donut (#48), Pie (#64), Stacked Bar (#38)** | **Bullet list** |
| **Content with visual / photo** | **Content+Right Image (#40), Left Image+Content (#41), Three Images (#42)** | **Text-only layouts** |
| **Risk / evaluation matrix** | **Risk Matrix (#54), SWOT (#65), Harvey Ball (#56), 2x2 Matrix (#13)** | **Bullet list** |
| **Strategic recommendations** | **Numbered List+Panel (#69), Decision Tree (#60), Checklist (#61)** | **Two-column text** |
| **Multi-KPI executive dashboard** | **Dashboard KPI+Chart (#57), Dashboard Table+Chart (#58)** | **Simple table** |
| **Stakeholder / relationship** | **Stakeholder Map (#59)** | **Bullet list** |
| **Meeting agenda** | **Agenda (#66)** | **Plain text** |

**NEVER** use Two-Column Text (#26) for more than 1 slide per deck. It is the least visually engaging layout.

**CHART PRIORITY RULE**: When data contains dates/periods + numeric values or percentages (e.g., `3/4 正面 20% 中性 80%` or `Q1: ¥850万`), you **MUST** use a Chart pattern (#37-#39, #48-#56, #64, #70) instead of a text-based layout. Charts maximize data-ink ratio and are the most visually compelling way to present time-series data.

**IMAGE PRIORITY RULE** (v1.8): When the content involves case studies, product showcases, location overviews, or any scenario where a visual/photo would strengthen the narrative, prefer Image+Content layouts (#40-#47, #68) over text-only layouts. The `add_image_placeholder()` function creates gray placeholder boxes that users replace with real images after generation.

### Content Density Requirements

"Minimalism" in McKinsey design means **removing decorative noise** (shadows, gradients, clip-art), NOT removing content. A slide with 80% whitespace is not minimalist — it is EMPTY.

**Mandatory minimums per content slide**:

1. **At least 3 distinct visual blocks** — e.g., title bar + content area + takeaway box, or title + left panel + right panel
2. **Body text area utilization ≥ 50%** of the available content space (between title bar at 1.4" and source line at 7.05")
3. **Action Title must be a FULL SENTENCE** expressing the slide's key insight:
   - ✅ `"连接组约束的AI模型将自由参数减少90%,实现单细胞精度预测"`
   - ❌ `"连接组约束的AI模型"`
4. **Use specific data points** when the user provides them (numbers, percentages, names) — display them prominently with Big Number or Metric Card patterns
5. **Source attribution** (`add_source()`) on every content slide with specific references, not generic labels

### Production Guard Rails (v1.9 / v2.0)

These rules address **recurring production defects** observed across multiple presentation generations. Each rule is derived from real-world user feedback and must be followed without exception.

#### Rule 1: Spacing Between Content Blocks and Bottom Bars

**Problem observed**: Tables, charts, or content grids placed immediately above a bottom summary/action bar (e.g., "行动公式", "趋势判读", "风险提示") with zero vertical gap, making them visually merged.

**MANDATORY**: There MUST be **at least 0.15" vertical gap** between the last content block and any bottom bar/summary box. Calculate positions explicitly:

```python
# ❌ WRONG: content ends at Inches(6.15), bottom bar also at Inches(6.15)
last_content_bottom = content_top + num_rows * row_height
bar_y = last_content_bottom  # NO GAP!

# ✅ CORRECT: explicit gap
BOTTOM_BAR_GAP = Inches(0.2)
bar_y = last_content_bottom + BOTTOM_BAR_GAP
```

**Validation formula**: `bottom_bar_y >= last_content_bottom + Inches(0.15)`

#### Rule 2: Content Overflow Protection

**Problem observed**: Text or shapes extending beyond the right margin (left_margin + content_width) or bottom margin (source line at 7.05").

**MANDATORY** overflow checks:

1. **Right margin**: Every element's `left + width ≤ LM + CW` (i.e., `Inches(0.8) + Inches(11.733) = Inches(12.533)`)
2. **Bottom margin**: Every element's `top + height ≤ Inches(6.95)` (leaving room for source line at 7.05")
3. **Text in bounded boxes**: When placing text inside a colored `add_rect()` box, the text box MUST be **inset by at least 0.15"** on each side:

```python
# ✅ CORRECT: text inset within its container box
box_left = LM
box_width = CW
add_rect(s, box_left, box_y, box_width, box_h, BG_GRAY)
add_text(s, box_left + Inches(0.3), box_y, box_width - Inches(0.6), box_h,
         text, ...)  # 0.3" padding on each side
```

4. **Multi-column layouts**: When calculating column widths, account for inter-column gaps AND the right margin:
   ```python
   # total available = CW = Inches(11.733)
   num_cols = 3
   gap = Inches(0.2)
   col_w = (CW - gap * (num_cols - 1)) / num_cols  # NOT CW / num_cols
   ```

5. **Long text truncation**: If generated text may exceed box boundaries, reduce `font_size` by 1-2pt or abbreviate text. Never allow visible overflow.

#### Rule 3: Bottom Whitespace Elimination

**Problem observed**: Charts or content areas end at ~Inches(5.5) while the bottom bar sits at ~Inches(6.3), leaving ~0.8" of dead whitespace.

**MANDATORY**: The bottom summary bar should be positioned at **no higher than Inches(6.1)** and **no lower than Inches(6.4)**. Adjust chart/content heights to fill available space. Target: visible whitespace between content and bottom bar ≤ 0.3".

```python
# ✅ CORRECT: Compute bottom bar position dynamically
content_bottom = chart_top + chart_height
# Place bottom bar close to content (but with minimum gap)
bar_y = max(content_bottom + Inches(0.15), Inches(6.1))
bar_y = min(bar_y, Inches(6.4))  # don't push past safe zone
```

#### Rule 4: Legend Color Consistency

**Problem observed**: Chart legends using plain black text "■" symbols (`■ 基准值 ■ 增加 ■ 减少`) while actual chart bars use NAVY, ACCENT_RED, ACCENT_GREEN — colors don't match.

**MANDATORY**: Every legend indicator MUST use a **colored square** (`add_rect()`) matching the exact color used in the chart below it. Never use text-only legends with "■" character.

```python
# ❌ WRONG: Text-only legend with black squares
add_text(s, LM, legend_y, CW, Inches(0.25),
         '■ 基准值  ■ 增加  ■ 减少', ...)

# ✅ CORRECT: Color-matched legend squares
lgx = LM + Inches(5)
add_rect(s, lgx, legend_y, Inches(0.15), Inches(0.15), NAVY)
add_text(s, lgx + Inches(0.2), legend_y, Inches(0.9), Inches(0.25),
         '基准值', font_size=Pt(10), font_color=MED_GRAY)
add_rect(s, lgx + Inches(1.3), legend_y, Inches(0.15), Inches(0.15), ACCENT_RED)
add_text(s, lgx + Inches(1.5), legend_y, Inches(0.9), Inches(0.25),
         '增加', font_size=Pt(10), font_color=MED_GRAY)
# ... repeat for each series
```

**Legend placement**: Inline with or immediately below the chart subtitle line (typically at Inches(1.15)-Inches(1.20)). Legend squares are 0.15" × 0.15" with 0.05" gap to label text.

#### Rule 5: Title Style Consistency

**Problem observed**: Some slides using `add_navy_title_bar()` (full-width navy background + white text) while others use `add_action_title()` (white background + black text + underline), creating jarring visual inconsistency.

**MANDATORY**: Use **`add_action_title()`** (`aat()`) as the **ONLY** title style for ALL content slides. The navy title bar (`antb()`) is **DEPRECATED for content slides** and should only appear if explicitly requested by the user.

```python
# ❌ DEPRECATED: Navy background title bar
def add_navy_title_bar(slide, text):
    add_rect(s, 0, 0, SW, Inches(0.75), NAVY)
    add_text(s, LM, 0, CW, Inches(0.75), text, font_color=WHITE, ...)

# ✅ CORRECT: Consistent white-background action title (bottom-anchored)
def add_action_title(slide, text, title_size=Pt(22)):
    add_text(s, Inches(0.8), Inches(0.15), Inches(11.7), Inches(0.9), text,
             font_size=title_size, font_color=BLACK, bold=True, font_name='Georgia',
             anchor=MSO_ANCHOR.BOTTOM)  # BOTTOM: text sits flush against separator
    add_hline(s, Inches(0.8), Inches(1.05), Inches(11.7), BLACK, Pt(0.5))
```

**Note**: When `add_action_title()` is used, content starts at **Inches(1.25)** (not Inches(1.0)). Account for this when positioning grids, tables, or charts below the title.

#### Rule 6: Axis Label Centering in Matrix/Grid Charts

**Problem observed**: In 2×2 matrix layouts (#13, #59, #65), axis labels ("用户规模↑", "技术壁垒→") positioned at fixed offsets rather than centered on their respective axes, causing visual misalignment.

**MANDATORY**: Axis labels MUST be **centered on the full span** of their axis:

```python
# Grid dimensions
grid_left = LM + Inches(2.0)
grid_top = Inches(1.65)
cell_w = Inches(4.5)  # width of each quadrant
cell_h = Inches(2.0)  # height of each quadrant
grid_w = 2 * cell_w   # full grid width
grid_h = 2 * cell_h   # full grid height

# ✅ CORRECT: Y-axis label centered vertically on FULL grid height
add_text(s, LM, grid_top, Inches(1.8), grid_h,
         'Y轴标签↑', alignment=PP_ALIGN.CENTER, anchor=MSO_ANCHOR.MIDDLE)

# ✅ CORRECT: X-axis label centered horizontally on FULL grid width
add_text(s, grid_left, grid_top + grid_h + Inches(0.1), grid_w, Inches(0.3),
         'X轴标签 →', alignment=PP_ALIGN.CENTER)
```

#### Rule 7: Image Placeholder Slide Requirement

**Problem observed**: Presentations generated with zero image-containing slides, resulting in a wall of text/charts that feels monotonous and lacks visual relief.

**MANDATORY**: For presentations with **8+ slides**, at least **1 slide** must include image placeholders (using `add_image_placeholder()` or custom gray boxes with "请插入图片" labels). Preferred positions:

- After the first 2-3 content slides (as a visual break)
- For case studies, product showcases, or ecosystem overviews

**Standard placeholder style** (when not using `add_image_placeholder()` helper):

```python
# Large placeholder
img_l = LM; img_t = Inches(1.3); img_w = Inches(6.5); img_h = Inches(4.0)
add_rect(s, img_l, img_t, img_w, img_h, BG_GRAY)
add_rect(s, img_l + Inches(0.04), img_t + Inches(0.04),
         img_w - Inches(0.08), img_h - Inches(0.08), WHITE)
add_rect(s, img_l + Inches(0.08), img_t + Inches(0.08),
         img_w - Inches(0.16), img_h - Inches(0.16), RGBColor(0xF8, 0xF8, 0xF8))
add_text(s, img_l, img_t + img_h // 2 - Inches(0.3), img_w, Inches(0.5),
         '[ 请插入图片 ]', font_size=Pt(22), font_color=LINE_GRAY,
         bold=True, alignment=PP_ALIGN.CENTER, anchor=MSO_ANCHOR.MIDDLE)
add_text(s, img_l, img_t + img_h // 2 + Inches(0.2), img_w, Inches(0.3),
         '图片描述标签', font_size=Pt(13), font_color=MED_GRAY,
         alignment=PP_ALIGN.CENTER, anchor=MSO_ANCHOR.MIDDLE)
```

This triple-border style (BG_GRAY → WHITE → #F8F8F8) creates a professional, clearly identifiable placeholder that prompts users to insert real images.

#### Rule 8: Dynamic Sizing for Variable-Count Layouts (v1.10.4)

**Problem observed**: Layouts with a variable number of items (checklist rows, value chain stages, cover multi-line titles) use **fixed dimensions** that only work for a specific count. When item count differs, content either overflows past page boundaries or leaves excessive whitespace.

**MANDATORY**: For any layout where the number of items is variable, compute dimensions dynamically:

```python
# ✅ Horizontal items (value chain, flow): fill content width
n = len(items)
gap = Inches(0.35)
item_w = (CW - gap * (n - 1)) / n   # NOT a fixed Inches(2.0)

# ✅ Vertical items (checklist, table rows): fit within available height
bottom_limit = BOTTOM_BAR_Y if bottom_bar else SOURCE_Y - Inches(0.05)
available_h = bottom_limit - content_start_y
item_h = min(MAX_ITEM_H, available_h / max(n, 1))  # cap at comfortable max

# ✅ Multi-line titles: height scales with line count
n_lines = text.count('\n') + 1
title_h = Inches(0.8 + 0.65 * max(n_lines - 1, 0))
# Position following elements relative to title bottom, NOT at fixed y
```

**Anti-patterns** (❌ NEVER DO):
- `stage_w = Inches(2.0)` for N stages → use `(CW - gap*(N-1)) / N`
- `row_h = Inches(0.55)` for N rows → use `min(0.85, available / N)`
- `subtitle_y = Inches(3.5)` on cover → use `title_y + title_h + Inches(0.3)`

#### Rule 9: BLOCK_ARC Native Shapes for Circular Charts (v2.0)

**Problem observed**: Donut charts (#48), pie charts (#64), and gauge dials (#55) rendered with hundreds to thousands of small `add_rect()` blocks. This creates 100-2800 shapes per chart, inflates file size by 60-80%, slows generation to 2+ minutes, and produces visual artifacts (gaps between blocks, jagged edges).

**MANDATORY**: Use **BLOCK_ARC** preset shapes via `python-pptx` + XML adjustment for all circular/arc charts. Each segment = 1 shape (total: 3-5 shapes per chart vs. hundreds).

**BLOCK_ARC angle convention** (PPT coordinate system):
- Angles measured **clockwise from 12 o'clock** (top), in **60000ths of a degree**
- Top = 0°, Right = 90°, Bottom = 180°, Left = 270°
- Example: a full-circle donut segment from 12 o'clock CW to 3 o'clock = adj1=0, adj2=5400000

**Three adj parameters**:
- `adj1`: start angle (60000ths of degree, CW from top)
- `adj2`: end angle (60000ths of degree, CW from top)
- `adj3`: inner radius ratio (0 = solid sector / pie, 50000 = max / invisible ring)

```python
from pptx.oxml.ns import qn

def add_block_arc(slide, left, top, width, height, start_deg, end_deg, inner_ratio, color):
    """Draw a BLOCK_ARC shape with precise angle and ring-width control.

    Args:
        slide: pptx slide object
        left, top, width, height: bounding box (width == height for circular arc)
        start_deg: start angle in degrees, CW from 12 o'clock (0=top, 90=right, 180=bottom, 270=left)
        end_deg: end angle in degrees, CW from 12 o'clock
        inner_ratio: 0 = solid pie sector, 50000 = max (paper-thin ring).
                     For ~10px ring width: int((outer_r - Pt(10)) / outer_r * 50000)
        color: RGBColor fill color
    """
    from pptx.enum.shapes import MSO_SHAPE
    sh = slide.shapes.add_shape(MSO_SHAPE.BLOCK_ARC, left, top, width, height)
    sh.fill.solid()
    sh.fill.fore_color.rgb = color
    sh.line.fill.background()
    _clean_shape(sh)  # remove p:style to prevent file corruption

    sp = sh._element.find(qn('p:spPr'))
    prstGeom = sp.find(qn('a:prstGeom'))
    if prstGeom is not None:
        avLst = prstGeom.find(qn('a:avLst'))
        if avLst is None:
            avLst = prstGeom.makeelement(qn('a:avLst'), {})
            prstGeom.append(avLst)
        for gd in avLst.findall(qn('a:gd')):
            avLst.remove(gd)
        gd1 = avLst.makeelement(qn('a:gd'), {'name': 'adj1', 'fmla': f'val {int(start_deg * 60000)}'})
        gd2 = avLst.makeelement(qn('a:gd'), {'name': 'adj2', 'fmla': f'val {int(end_deg * 60000)}'})
        gd3 = avLst.makeelement(qn('a:gd'), {'name': 'adj3', 'fmla': f'val {inner_ratio}'})
        avLst.append(gd1)
        avLst.append(gd2)
        avLst.append(gd3)
    return sh
```

**Usage patterns**:

```python
# ── Donut chart: 4 segments, ~10px ring width ──
outer_r = Inches(1.6)
inner_ratio = int((outer_r - Pt(10)) / outer_r * 50000)  # ~10px ring
cum_deg = 0  # start at top (0° = 12 o'clock)
for pct, color, label in segments:
    sweep = pct * 360
    add_block_arc(s, cx - outer_r, cy - outer_r, outer_r * 2, outer_r * 2,
                  cum_deg, cum_deg + sweep, inner_ratio, color)
    cum_deg += sweep

# ── Pie chart (solid sectors): inner_ratio = 0 ──
add_block_arc(s, cx - r, cy - r, r * 2, r * 2, 0, 151.2, 0, NAVY)  # 42%

# ── Horizontal rainbow gauge (semi-circle, left→top→right) ──
# PPT coords: left=270°, top=0°, right=90°
gauge_segs = [(0.40, ACCENT_RED), (0.30, ACCENT_ORANGE), (0.30, ACCENT_GREEN)]
inner_ratio = int((outer_r - Pt(10)) / outer_r * 50000)
ppt_cum = 270  # start at left
for pct, color in gauge_segs:
    sweep = pct * 180
    add_block_arc(s, cx - outer_r, cy - outer_r, outer_r * 2, outer_r * 2,
                  ppt_cum % 360, (ppt_cum + sweep) % 360, inner_ratio, color)
    ppt_cum += sweep
```

**Anti-patterns** (❌ NEVER DO for circular charts):
- Nested `for deg in range(...): for r in range(...): add_rect(...)` — generates hundreds/thousands of tiny squares
- Drawing a white circle on top of a filled circle to "fake" a donut — fragile, misaligns on resize
- Using `math.cos/sin` + `add_rect()` loops for arcs — always use `BLOCK_ARC` instead

### Mandatory Slide Elements

EVERY content slide (except Cover and Closing) MUST include ALL of these:

| Element | Function | Position |
|---------|----------|----------|
| Action Title | `add_action_title(slide, text)` | Top (0.15" from top) |
| Title separator line | Included in `add_action_title()` | 1.05" from top |
| Content area | Layout-specific content blocks | 1.4" to 6.5" |
| Source attribution | `add_source(slide, text)` | 7.05" from top |
| Page number | `add_page_number(slide, n, total)` | Bottom-right corner |

Page number helper function:
```python
def add_page_number(slide, num, total):
    add_text(slide, Inches(12.2), Inches(7.1), Inches(1), Inches(0.3),
             f"{num}/{total}", font_size=Pt(9), font_color=MED_GRAY,
             alignment=PP_ALIGN.RIGHT)
```

---

## Layout Patterns

### Slide Dimensions

```python
prs.slide_width = Inches(13.333)
prs.slide_height = Inches(7.5)
```

Widescreen format (16:9), standard for all presentations.

### Standard Margin/Padding

| Position | Size | Usage |
|----------|------|-------|
| **Left margin** | 0.8" | Default left edge |
| **Right margin** | 0.8" | Default right edge |
| **Top (below title)** | 1.4" | Content start position |
| **Bottom** | 7.05" | Source text baseline |
| **Title bar height** | 0.9" | Action title bar |
| **Title bar top** | 0.15" | From slide top |

### Slide Type Patterns

#### 1. Cover Slide (Slide 1)

Layout:
- Navy bar at very top (0.05" height)
- Main title (44pt, Georgia, navy) at y=1.2" — **height computed dynamically from line count**
- Subtitle (24pt, dark gray) positioned **below title dynamically**
- Date/info (14pt, med gray) follows subtitle
- Decorative navy line at x=1", y=6.8" (4" wide, 2pt)

Code template:
```python
s1 = prs.slides.add_slide(prs.slide_layouts[6])
add_rect(s1, 0, 0, prs.slide_width, Inches(0.05), NAVY)

# Dynamic title height based on line count
lines = title.split('\n') if isinstance(title, str) else title
n_lines = len(lines) if isinstance(lines, list) else title.count('\n') + 1
title_h = Inches(0.8 + 0.65 * max(n_lines - 1, 0))

add_text(s1, Inches(1), Inches(1.2), Inches(11), title_h,
         '项目名称', font_size=Pt(44), font_name='Georgia',
         font_color=NAVY, bold=True, ea_font='KaiTi')

# Position elements BELOW title dynamically — never use fixed y
sub_y = Inches(1.2) + title_h + Inches(0.3)
add_text(s1, Inches(1), sub_y, Inches(11), Inches(0.8),
         '副标题描述', font_size=Pt(24),
         font_color=DARK_GRAY, ea_font='KaiTi')
sub_y += Inches(1.0)

add_text(s1, Inches(1), sub_y + Inches(0.3), Inches(11), Inches(0.5),
         '演示文稿  |  2026年3月', font_size=BODY_SIZE,
         font_color=MED_GRAY, ea_font='KaiTi')
add_hline(s1, Inches(1), Inches(6.8), Inches(4), NAVY, Pt(2))
```

#### 2. Action Title Slide (Most Content Slides)

Every main content slide has this structure:

```
┌─────────────────────────────────────────┐ 0.15"
│ ▌ Action Title (22pt, bold, black)      │ ← TITLE_BAR_H = 0.9"
├─────────────────────────────────────────┤ 1.05"
│                                         │
│  Content area (starts at 1.4")          │
│  [Tables, lists, text, etc.]            │
│                                         │
│                                         │
│  ──────────────────────────────────────  │ 7.05"
│  Source: ...                            │ 9pt, med gray
└─────────────────────────────────────────┘ 7.5"
```

Code pattern:
```python
s = prs.slides.add_slide(prs.slide_layouts[6])
add_action_title(s, 'Slide Title Here')
# Then add content below y=1.4"
add_source(s, 'Source attribution')
```

#### 3. Table Layout (Slide 4 - Five Capabilities)

Structure:
- Header row with column names (BODY_SIZE, gray, bold)
- 1.0pt black line under header
- Data rows (1.0" height each, 14pt text)
- 0.5pt gray line between rows
- 3 columns: Module (1.6" wide), Function (5.0"), Scene (5.1")

```python
# Headers
add_text(s4, left, top, width, height, text,
         font_size=BODY_SIZE, font_color=MED_GRAY, bold=True)

# Header line (thicker)
add_line(s4, left, top + Inches(0.5), left + full_width, top + Inches(0.5),
         color=BLACK, width=Pt(1.0))

# Rows
for i, (col1, col2, col3) in enumerate(rows):
    y = header_y + row_height * i
    add_text(s4, left, y, col1_w, row_h, col1, ...)
    add_text(s4, left + col1_w, y, col2_w, row_h, col2, ...)
    add_text(s4, left + col1_w + col2_w, y, col3_w, row_h, col3, ...)
    # Row separator
    add_line(s4, left, y + row_h, left + full_w, y + row_h,
             color=LINE_GRAY, width=Pt(0.5))
```

#### 4. Three-Column Overview (Slide 5)

Layout:
- Left column (4.1" wide): "是什么"
- Middle column (4.1" wide): "独到之处"
- Right 1/4 (2.5" wide) gray panel: "Key Takeaways"

```
0.8"  4.9"  5.3"  9.4"  10.0" 12.5"
|-----|-----|-----|-----|------|
│左 │ │ 中 │ │ 右 │
└─────────────────────────────┘
```

Code:
```python
left_x = Inches(0.8)
col_w5 = Inches(4.1)
mid_x = Inches(5.3)
takeaway_left = Inches(10.0)
takeaway_width = Inches(2.5)

# Left column
add_text(s5, left_x, content_top, col_w5, ...)
add_text(s5, left_x, content_top + Inches(0.6), col_w5, ..., 
              bullet=True, line_spacing=Pt(8))

# Right gray takeaway area
add_rect(s5, takeaway_left, Inches(1.2), takeaway_width, Inches(5.6), BG_GRAY)
add_text(s5, takeaway_left + Inches(0.15), Inches(1.35), takeaway_width - Inches(0.3), ...,
         'Key Takeaways', font_size=BODY_SIZE, color=NAVY, bold=True)
add_text(s5, takeaway_left + Inches(0.15), Inches(1.9), takeaway_width - Inches(0.3), ...,
              [f'{i+1}. {t}' for i, t in enumerate(takeaways)], line_spacing=Pt(10))
```

---

### 类别 A:结构导航

#### 5. Section Divider (章节分隔页)

**适用场景**: 多章节演示文稿的章节过渡页,用于视觉上分隔不同主题模块。

```
┌──┬──────────────────────────────────────┐
│N │                                      │
│A │  第一部分                             │
│V │  章节标题(28pt, NAVY, bold)          │
│Y │  副标题说明文字                        │
│  │                                      │
└──┴──────────────────────────────────────┘
```

```python
s = prs.slides.add_slide(BL)
add_rect(s, 0, 0, Inches(0.6), SH, NAVY)
add_text(s, Inches(1.2), Inches(2.0), Inches(10), Inches(0.8),
         '第一部分', font_size=SUB_HEADER_SIZE, font_color=MED_GRAY, font_name='Georgia')
add_text(s, Inches(1.2), Inches(2.8), Inches(10), Inches(1.2),
         '章节标题', font_size=HEADER_SIZE, font_color=NAVY, bold=True, font_name='Georgia')
add_text(s, Inches(1.2), Inches(4.2), Inches(10), Inches(0.6),
         '副标题说明文字', font_size=BODY_SIZE, font_color=DARK_GRAY)
```

#### 6. Table of Contents / Agenda (目录/议程页)

**适用场景**: 演示文稿开头的目录或会议议程,列出各章节及说明。

```
┌─────────────────────────────────────────┐
│ ▌ 目录                                  │
├─────────────────────────────────────────┤
│                                         │
│  (1)  章节一标题     简要描述            │
│  ─────────────────────────────────────  │
│  (2)  章节二标题     简要描述            │
│  ─────────────────────────────────────  │
│  (3)  章节三标题     简要描述            │
│                                         │
└─────────────────────────────────────────┘
```

```python
s = prs.slides.add_slide(BL)
add_action_title(s, '目录')
items = [('1', '引言与背景', '项目起源与核心问题'),
         ('2', '市场分析', '竞争格局与机会识别'),
         ('3', '战略建议', '三大核心行动方案')]
iy = Inches(1.6)
for num, title, desc in items:
    add_oval(s, LM, iy, num, size=Inches(0.5))
    add_text(s, LM + Inches(0.7), iy, Inches(4.0), Inches(0.4),
             title, font_size=SUB_HEADER_SIZE, font_color=NAVY, bold=True)
    add_text(s, In

This file is long; only the first 41 KB is shown here. See the source page for the full text.

Popular in Office

腾讯文档 TENCENT DOCSOffice

A full manual for creating, editing, and managing Tencent Docs through MCP.

↓ 2.2M★ 327u_b0de8114
WPS Office 全家桶Office

WPS Office 全家桶 - 四引擎(WPS/MS Office/LibreOffice/纯Python)智能识别用户已安装软件,纯Python模式支持排序/筛选/图表/公式/统计,含文档模板(代码生成)、最佳实践案例、故障排除大章(20+避坑+15 FAQ+15错误ID 统一索引)、自动重试、硬件自适应、环境自检、Skill更新提醒;v5.1新增:kingdoc云端桥接(白名单探测+配置预检+subprocess上传+差异检查)、条件格式/数据验证跨引擎IO(JSON中间格式保真往返)、会议纪要三段式要素化(待办清单+决策记录+风险异议)+说话人标注(MFCC聚类);v5.2新增:公文一键排版(GB/T 9704 规则库约束+结构识别套版+红头预留位+联合行文)、批量水印与敏感信息打码(身份证/手机号/银行卡/邮箱双轨识别+黑色矩形覆盖+打码清单)、跨文档知识检索与问答(本地TF-IDF零依赖+跳转+可选zwjh记忆桥接);删除:与contract-review重复的轻量合同审查模块,改为文档互链引导;v5.2.5修复:MS Office引擎下Excel/PPT创建必然失败(错用Word COM对象),新增get_ms_excel/get_ms_ppt及三槽位统一释放release_ms,MS Office路径透传filepath参数;v5.3.5修复:SKILL.md frontmatter 修复 tags 块序列误混入联系方式行导致标准 YAML 解析失败;v5.4.0新增:留存型功能——定时任务与报告中心(schedule_register 升级为三平台统一调度层 schtasks/cron/launchd,错过策略 skip/run_once,注册表 JSON 可导出导入;新建 report_center.py 报告中心,templates/reports/*.json 参数化模板 + ECharts 单文件 HTML 输出(图表内联、无服务端、双击即开);report_generator 升级支持 --format html;wps.py CLI 统一埋点 SQLite,自动供 report_center 生成周度使用摘要附在报告末尾;watch.py 职责收窄为纯事件采集);v5.3.0新增:统一CLI入口 wps(registry.json注册表分发,office --app/--action 与 convert 子命令、--list 一屏列出全部命令,薄路由层不改脚本内部逻辑)、全命令 --json 机器可读输出(固定 ok/error_id/elapsed/data 四字段,对齐 E001-E015,wps_error.py 新增 emit_json 统一出口)、_worker 启动优化(wps_pure 改惰性加载,pywin32/python-docx/openpyxl 经核查已惰性导入,ENGINE_CACHE 进程级单例复用,COM 释放统一走 release_ms 三槽位)、转换三入口合并为 convert(format_converter/md_converter/wps_docx_to_ppt 内部按目标格式路由,旧脚本保留为兼容别名并提示新命令,v6.0.0 移除)

↓ 766.6K★ 159user_1a470ba8
ppt-generator-skillOffice

A skill that generates PPT decks from a topic and style.

↓ 592.7K★ 346user_3b34947d
秘书材料星公文写作-妙笔skillOffice

体制内资深办公室主任+管理学博士,自学编程开发的智能公文写作Skill,让所有办公室同仁少掉头发、少加无用的班,材料一稿过关,省下的时间留给自己、享受生活!————————我们能给你带来的核心价值:1、颠覆传统写作:逻辑建模38类公文全链条,特有模板化引导问答模式,分分钟省心生成高质量大稿。--------2、替你“向上管理”:独有审稿人画像机制,文稿一键结构性修改,精准匹配审稿人画像,过稿率提升至少10倍!----------3、‌杜绝AI编造:缺失的数字、文号、人名、日期一律用[待补]标注,绝不杜撰,守住公文底线。-----------4、‌算法格式托底:数十万全量智控知识库+6500行前沿核心控制代码,格式精准对齐GB/T 9704,出稿零失误。——————————————————本skill历史版本开放了5%知识库的免费试用版,下载即可使用,但需注意:知识库有著作权,仅用于个人使用,不可用于其他任何目的。有兴趣的可以试试。——————————————————————单次付费版skill已发布,2元一次使用全量知识库。具体在发布者“熵衍信息”主页skill里。

↓ 578.4K★ 46u_70fc650d
WPS Office 全家桶Office

WPS Office 全家桶 - 四引擎(WPS/MS Office/LibreOffice/纯Python)智能识别用户已安装软件,纯Python模式支持排序/筛选/图表/公式/统计,含文档模板(代码生成)、最佳实践案例、故障排除大章(20+避坑+15 FAQ+15错误ID 统一索引)、自动重试、硬件自适应、环境自检、Skill更新提醒;v5.0新增:定时任务注册管理(schtasks/crontab)、目录轮询监听(YAML规则+大文件分片)、使用统计与稳定性中心(SQLite埋点+引擎健康+引擎偏好)、COM三步自愈(regsvr32→修复安装→手动指引)、对话式数据查询(NL2SQL式意图解析+查询计划+连续追问)、multi-agent编排引擎桥接(白名单探测+流水线模板注册)

↓ 508.0K★ 4user_814dbe54
WPS全家桶·办公管家.SkillOffice

WPS/MS Office/LibreOffice/纯Python 四引擎智能办公全家桶:文档/表格/PPT/公文/会议纪要一体化。

↓ 429.5K★ 2user_185644ed
腾讯文档在线协作自动化Office

通过 Tencent Docs Open API 以 curl + jq 创建、读取、列出、搜索和管理腾讯文档的自动化工作流,覆盖在线文档、在线表格、幻灯片、思维导图、流程图、智能表格与收集表七种类型,支持读取元数据与单元格区域、上传图片、创建副本,并对重命名/移动/删除等写操作强制先经用户确认。Use when the user 提到腾讯文档 / Tencent Docs / docs.qq.com、粘贴 docs.qq.com 链接,或想在其腾讯文档空间中创建、读取、整理文档时使用。Triggers include 腾讯文档、在线文档、在线表格、思维导图、流程图、智能表格、收集表。如需文案写作完整工作流套件,淘宝搜索「AI灵匣」店铺获取完整版。

↓ 428.1K★ -u_1ea983ac
kdocs skillOffice

The official skill for working with Kdocs cloud documents through kdocs-cli.

↓ 424.1K★ 128user_5ea84866

Data as of 2026-10. Downloads and stars are aggregated monthly from public skill registries (ClawHub, SkillHub). See data methodology.