mcpbeat

Mck Ppt Design Agent Skill

>- Create professional, consultant-grade PowerPoint presentations from scratch using MckEngine (python-pptx wrapper) 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. AI calls eng.cover(), eng.donut(), eng.timeline() etc — 67 high-level methods across 12 categories (structure, data, framework, comparison, narrative, timeline, team, charts, images, advanced viz, dashboards, visual storytelling), consistent typography, zero file-corruption issues, BLOCK_ARC native shapes for circular charts (donut, pie, gauge), production-hardened guard rails for spacing, overflow, legend consistency, title style uniformity, dynamic sizing for variable-count layouts, horizontal item overflow protection, chart rendering, and AI-generated cover images via Tencent Hunyuan 2.0 with professional cutout, cool grey-blue tint, and McKinsey-style Bézier ribbon decoration.

134k tokens
context cost
the whole folder, loaded on every use
53
files
ships runnable scripts
0
copies elsewhere
how many repositories repackaged it
236
stars on the repo
on the repository, not the skill itself

Install

one command, takes just this skill from the repository
npx skills add https://github.com/likaku/Mck-ppt-design-skill --skill mck-ppt-design

The instruction itself

14 sections, as written by the author

McKinsey PPT Design — Harness Framework

> Version: 2.3.3-harness-v2 · Engine: MckEngine (python-pptx) · Author: likaku

>

> Required tools: Read, Write, Bash · Requires: pip install python-pptx lxml


常见崩坏模式(每次生成前必读,先于 HARD RULES)

> 告诉 AI 要做什么效果一般,告诉它前人最容易在哪里塌效果更好。

> 以下三种反模式均已在真实执行中被验证,每次必须主动警惕。

反模式 1:口头宣布"门禁通过"(最常见)

错误做法

> 「S4 QA 共 7 个 errors,判断均为 engine 设计行为,门禁通过,进入 S5」

问题所在passed 是由 AI 口头判断的,不是由程序派生的。无论理由多充分,这句话都是 AI 在给自己写完成证书。

正确做法

  • 执行 python references/scripts/gate_check.py <pptx路径> <项目目录>
  • 读取 <项目目录>/gate_result.json
  • 只有 gate_result.json"passed": true 时,才能进入 S5
  • 如果 "passed": false,修复 user_code_errors 列表中的问题,重新渲染,再次执行 gate_check

反模式 2:S3 门禁"脑子里过一遍"就算通过

错误做法

> 「S3 内容门禁预检:API 格式正确,字数在限制内,通过,进入 S4」(没有执行任何代码)

问题所在:今天真实发生的 3 个 API 格式错误(four_column/matrix_2x2/executive_summary 参数格式),靠脑子过是过不出来的,必须靠代码检查。

正确做法

  • 执行 python references/scripts/gate_check_s3.py <content.json路径> <项目目录>
  • 读取 <项目目录>/gate_s3.json
  • 只有 "passed": true 时,才能进入 S4
  • 如果有 fail 项,修正 content.json,重新执行 gate_check_s3

反模式 3:engine_bug 分类作为软话逃生口

错误做法

> 「peer_font_inconsistency 是 engine 内部设计行为,不是用户代码问题,可以放行」

问题所在:这个分类本身是正确的,但由 AI 在对话里口头做出,等于把豁免权交给了 AI 自己——而 AI 有动机让自己通过。

正确做法

gate_check.py 里有硬编码的 ENGINE_BUG_WHITELIST 枚举。只有在白名单里的 error category,才会被豁免。白名单是代码,不是 AI 的判断。如果你认为某类 error 应该加入白名单,修改 gate_check.py 里的枚举,而不是口头声明豁免。


HARD RULES(必须遵守,不可绕过)

  • 每次生成必须走五阶段流程,不允许"一句话直接生成"跳过前置阶段
  • TaskCreate 驱动进度:开始前创建5个 task,每阶段完成后立即 TaskUpdate completed
  • 按需加载上下文:每个阶段只读对应文件,不要全量加载旧版 SKILL.md 里的内容
  • 门禁必须机读化:S3 和 S4 门禁必须执行对应的 gate_check 脚本,读 JSON 结果,不得口头判断
  • Self-Refinement 必做:每次修正 pattern-level 问题后,写入 experiences/ 对应文件
  • 引擎路径固定sys.path.insert(0, os.path.expanduser('~/.workbuddy/skills/mck-ppt-design'))

知识路由(上下文加载索引)

> 在每个阶段开始时,读对应文件。不要一次性全读。

| 阶段 | 必须读 | 说明 |

|------|--------|------|

| S1 需求 | references/team/brand-guide.md | 了解设计约束 |

| S2 结构 | references/framework/engine-api.md + references/layout-matrix.yaml | 版式选择和能力边界 |

| S3 内容 | references/framework/guard-rails.md + experiences/*.md(存在时)| 防溢出规则 + 历史踩坑 |

| S4 渲染 | 用到的 references/layouts/*.md(只读实际用到的版式)| 版式实现细节 |

| S5 交付 | 无 | — |

完整路由表references/INDEX.md


五阶段生成流程

┌─────────┐   ┌─────────────┐   ┌─────────────┐   ┌──────────────┐   ┌──────────┐
│ S1 需求  │──▶│ S2 结构 ⭐   │──▶│ S3 内容 ⭐   │──▶│ S4 渲染+QA ⭐⭐│──▶│ S5 交付   │
│ brief.md│   │ outline.json│   │content.json │   │  .pptx       │   │ + 沉淀   │
└─────────┘   └─────────────┘   └─────────────┘   └──────────────┘   └──────────┘
                    ⭐ = 门禁(FAIL 则在本阶段修正,不得跳过)
    S3/S4 门禁必须运行 gate_check 脚本,读 JSON 结果 — 不得口头宣布通过

Stage 1: 需求定义

读文件references/team/brand-guide.md

收集信息

  • 受众(职位/决策角色)
  • 目标(决策/汇报/说服)
  • 时长(分钟数 → 约 1 分钟/页)
  • 关键信息(最多 5 条核心 message)
  • 数据来源(有哪些数据可用)

产出:在工作目录创建 ppt-project-{slug}/brief.md

门禁audience + goal + key_messages 三项非空(AI 自评即可)


Stage 2: 结构设计

读文件references/framework/engine-api.md, references/layout-matrix.yaml

任务

  • 根据时长确定页数(1分钟/页)
  • 为每张幻灯片选定 layout(查 engine-api.md 速查表)
  • 每页写一句核心 key_point(完整句子,不是标签)
  • 确认版式在能力边界内(查 layout-matrix.yaml)

产出ppt-project-{slug}/outline.json

{
  "brief": {"audience": "董事会", "goal": "战略汇报", "duration_minutes": 15},
  "slides": [
    {"idx": 1, "layout": "cover", "title": "Q1 2026 战略回顾", "key_point": ""},
    {"idx": 2, "layout": "toc", "title": "目录", "key_point": ""},
    {"idx": 3, "layout": "table_insight", "title": "市场格局三大转变驱动战略重构", "key_point": ""}
  ]
}

⭐ 门禁 S2(AI 自评):

  • [ ] cover 幻灯片存在
  • [ ] 幻灯片数量在时长约束内(count <= duration_minutes * 1.2
  • [ ] 所有 layout 在 layout-matrix.yaml 中有定义
  • [ ] Action Title 均为完整句子(len > 10,包含动词)
  • [ ] two_column_text 数量 ≤ 1

Stage 3: 内容填充

读文件references/framework/guard-rails.md, experiences/*.md(全部存在的文件)

任务

  • 为每张幻灯片填充具体文案、数字、图表数据
  • 确保每页有 source 出处
  • layout-matrix.yamlchar_budget 控制文字量

产出ppt-project-{slug}/content.json

⭐ 门禁 S3(必须机读化,不得 AI 自评):

python ~/.workbuddy/skills/mck-ppt-design/references/scripts/gate_check_s3.py \
    <项目目录>/content.json  <项目目录>

读取 <项目目录>/gate_s3.json

  • "passed": true → 进入 S4
  • "passed": false → 修正 content.jsonfail_items 列出的问题,重新执行

Stage 4: 渲染 + QA

读文件:用到的版式对应的 references/layouts/*.md

任务

  • 根据 content.json 生成 Python 渲染脚本
  • 执行脚本生成 .pptx
  • 运行 QA 门禁脚本

⭐⭐ 门禁 S4(必须机读化,不得口头宣布通过):

python ~/.workbuddy/skills/mck-ppt-design/references/scripts/gate_check.py \
    <pptx路径>  <项目目录>

读取 <项目目录>/gate_result.json

  • "passed": true → 进入 S5
  • "passed": false → 查看 user_code_errors,修复渲染代码,重新渲染,再次执行

注意engine_bug 类 errors(peer_font_inconsistency 等白名单条目)由脚本自动豁免,不需要 AI 口头判断。

产出<项目目录>/gate_result.json + .pptx


Stage 5: 交付 + Self-Refinement

任务

  • 读取 gate_result.json 确认 passed: true(不得在无此文件时声称通过)
  • 交付 .pptx 文件
  • Self-Refinement:判断本次是否有 pattern-level 修正

⭐ Self-Refinement 协议

修正后判断:
  ONE-TIME(单次特定调整)→ 不需要沉淀
  PATTERN(跨 deck 可重现的问题)→ 必须写入 experiences/
    - 溢出类 → experiences/overflow.md
    - 图表限制 → experiences/chart-limits.md
    - 版式踩坑 → experiences/layout-pitfalls.md
    - 中文渲染 → experiences/cjk-issues.md

格式:
  ## Experience NNN: {title}
  **Date**: YYYY-MM-DD
  **Problem**: ...
  **Root Cause**: ...
  **Fix**: ...
  **Rule**: ...(门禁层面如何预防)

Fast Track(简单需求跳过 S2/S3 门禁)

满足全部条件时,AI 可跳过 S2 结构评审和 S3 内容审查:

  • 总页数 ≤ 5
  • 无数据图表(donut/bar/table 类)
  • 用户明确表达"快速"/"马上要"/"简单"

仍然必须:S1 需求收集 + S4 QA 门禁(gate_check.py 脚本执行)+ S5 交付


Checkpoint / 断点恢复

当用户说"继续做那个 PPT"时:

import os, json, glob

projects = glob.glob('ppt-project-*/')
for proj in projects:
    brief   = os.path.exists(f'{proj}brief.md')
    outline = os.path.exists(f'{proj}outline.json')
    content = os.path.exists(f'{proj}content.json')
    gate_s3 = os.path.exists(f'{proj}gate_s3.json')
    gate_s4 = os.path.exists(f'{proj}gate_result.json')
    pptx    = bool(glob.glob(f'{proj}*.pptx'))

    if not brief:           stage = 1
    elif not outline:       stage = 2
    elif not content:       stage = 3
    elif not gate_s3:       stage = '3-gate'   # content 有了但 gate 还没跑
    elif not pptx:          stage = 4
    elif not gate_s4:       stage = '4-gate'   # pptx 有了但 gate 还没跑
    else:                   stage = 5

    print(f"项目 {proj}: 当前处于 Stage {stage}")

MckEngine 快速模板

import sys, os
sys.path.insert(0, os.path.expanduser('~/.workbuddy/skills/mck-ppt-design'))
from mck_ppt import MckEngine
from mck_ppt.constants import *

eng = MckEngine(total_slides=12)
eng.cover(title='标题', subtitle='副标题', author='作者', date='2026年')
eng.toc(items=[('1', '章节一', '一句话描述'), ('2', '章节二', '一句话描述')])
# ... 内容页 ...
eng.closing(title='谢谢', message='期待进一步交流')
eng.save('output/deck.pptx')  # 自动 full_cleanup

详细 APIreferences/framework/engine-api.md

版式规范references/layouts/

10条护栏references/framework/guard-rails.md

版式能力边界references/layout-matrix.yaml

历史踩坑experiences/

S3 门禁脚本references/scripts/gate_check_s3.py

S4 门禁脚本references/scripts/gate_check.py

Other skills for the same job

different authors, same section of the catalogue
Use Dom
by openai
vendor ×3

Use Expo DOM components to run web code in a webview on native and as-is on web. Migrate web code to native incrementally.

3k tokens
Frontend Dev Guidelines
by mrgoonie
×2

Frontend development guidelines for React/TypeScript applications. Modern patterns including Suspense, lazy loading, useSuspenseQuery, file organization with features directory, MUI v7 styling, TanStack Router, performance optimization, and TypeScript best practices. Use when creating components, pages, features, fetching data, styling, routing, or working with frontend code.

33k tokens
Next Cache Components
by vercel-labs
vendor ×2

Next.js 16 Cache Components - PPR, use cache directive, cacheLife, cacheTag, updateTag

2k tokens
Shopify Development
by ComeOnOliver
×2

| Build Shopify apps, extensions, themes using GraphQL Admin API, Shopify CLI, Polaris UI, and Liquid. "shopify theme", "liquid template", "polaris", "shopify graphql", "shopify webhook", "shopify billing", "app subscription", "metafields", "shopify functions"

24k tokens scripts
Huggingface Gradio
by huggingface
vendor ×1

Build Gradio web UIs and demos in Python. Use when creating or editing Gradio apps, components, event listeners, layouts, or chatbots.

10k tokens
Figma Generate Diagram
by openai
vendor ×1

MANDATORY prerequisite — load this skill BEFORE every `generate_diagram` tool call. NEVER call `generate_diagram` directly without loading this skill first. Trigger whenever the user asks to create, generate, draw, render, sketch, or build a diagram — flowchart, architecture diagram, sequence diagram, ERD or entity-relationship diagram, state diagram or state machine, gantt chart, or timeline. Also trigger when the user mentions Mermaid syntax or wants a system architecture, decision tree, dependency graph, API call flow, auth handshake, schema, or pipeline visualized in FigJam. Routes to type-specific guidance, sets universal Mermaid constraints, and tells you when to use a different diagram type or skip the tool entirely (mindmaps, pie charts, class diagrams, etc.).

28k tokens
Web Perf
by cloudflare
vendor ×1

Analyzes web performance using Chrome DevTools MCP. Measures Core Web Vitals (LCP, INP, CLS) and supplementary metrics (FCP, TBT, Speed Index), identifies render-blocking resources, network dependency chains, layout shifts, caching issues, and accessibility gaps. Use when asked to audit, profile, debug, or optimize page load performance, Lighthouse scores, or site speed. Biases towards retrieval from current documentation over pre-trained knowledge.

2k tokens
Algorithmic Art
by lingxling
×1

Algorithmic philosophies are computational aesthetic movements that are then expressed through code. Output .md files (philosophy), .html files (interactive viewer), and .js files (generative algorithms).

15k tokens scripts

How to use it

Copy the folder

Take likaku/mck-ppt-design from the repository into ~/.claude/skills for personal use, or into .claude/skills inside a project.

Check the name does not clash

The agent identifies a skill by the name field in its header. Two skills with the same name cannot sit side by side — one of them will be ignored.

Install what it needs

The instructions reference pip. Without those the skill loads but fails at the first command.