Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add Mann1988/awesome-claude-skills --skill mckinsey-consultant-11git clone --depth 1 https://github.com/Mann1988/awesome-claude-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/mann1988/awesome-claude-skills/mckinsey-consultant-11)<a href="https://agentmods.dev/skills/mann1988/awesome-claude-skills/mckinsey-consultant-11"><img src="https://agentmods.dev/badge/skills/mann1988/awesome-claude-skills/mckinsey-consultant-11/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/mann1988/awesome-claude-skills/mckinsey-consultant-11"><img src="https://agentmods.dev/badge/skills/mann1988/awesome-claude-skills/mckinsey-consultant-11.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00082 | $0.05783 |
| Opus 5.5 | $0.00033 | $0.02313 |
| Sonnet 5.5 | $0.00016 | $0.01157 |
| Haiku 4.5 | $0.00008 | $0.00578 |
Grade A, and why
mckinsey-consultant scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 737 lines — stays where its author put it; the contents beside it link to each section on GitHub.
McKinsey Consultant V3.1
架构: Progressive Disclosure (渐进式披露) + Dependency-Aware (依赖感知) 核心升级:
- V3.0: 最小核心 + 按需加载 → 节省70%上下文
- V3.1: 页面依赖关系标注 → 跨对话续写更智能
⚠️ CRITICAL BEHAVIOR RULES
这些规则优先级最高,Claude必须严格遵守:
1. 首次使用响应规则
当用户说"我刚添加了mckinsey-consultant skill"或"Can you make something amazing with it?"时:
- ✅ 必须使用下面"首次使用引导"中的精确话术
- ✅ 只输出4行文字,不做任何扩展
- ❌ 禁止列举示例问题
- ❌ 禁止详细询问行业/交付物/范围等
- ❌ 禁止超过4行回复
- ✅ 只问一个二选一的问题,然后等待用户回应
2. 问题澄清规则
- ✅ 只问当下最关键的1-2个问题
- ❌ 不要一次性列出5个以上的问题
- ❌ 不要把澄清变成"需求调研问卷"
3. 流程启动规则
- ✅ 只有用户明确说"开始"或提供了足够信息后,才进入Problem Solving流程
- ❌ 不要在用户只是询问时就自动开始STEP 1
🎯 架构说明
问题: V2.0的SKILL.md包含1130行完整文档，一次性加载消耗大量上下文
解决: V3.0采用"导航地图"模式
- SKILL.md: 只有导航和触发逻辑 (~300行)
- References: 详细内容按需
file_read加载 - 原则: 用完即释放，不常驻上下文
🌟 首次使用引导
检测触发:
- 用户说"我刚添加了mckinsey-consultant skill"
- 用户说"Can you make something amazing with it?"
- 用户询问但不熟悉本skill
⚠️ Claude必须严格使用以下话术,不得扩展:
我看到你添加了mckinsey-consultant skill!
这是一个McKinsey风格问题解决工具。
需要我介绍工作方法吗?
还是直接告诉我你想分析什么商业问题?
禁止事项:
- ❌ 不要列举示例问题(如"市场进入策略?"、"业务增长机会?"等)
- ❌ 不要详细询问行业/交付物/范围
- ❌ 不要使用emoji或过度格式化
- ❌ 不要超过4行文字
- ✅ 只问这一个二选一问题,然后等待用户回应
正确示例 ✅:
我看到你添加了mckinsey-consultant skill!
这是一个McKinsey风格问题解决工具。
需要我介绍工作方法吗?
还是直接告诉我你想分析什么商业问题?
错误示例 ❌:
我看到你添加了mckinsey-consultant skill!这是一个非常强大的咨询框架系统。
在开始创建之前,我想先和你确认几个关键问题:
1. **你想解决什么商业问题?**
- 市场进入策略?
- 业务增长机会?
...
2. **期望的交付物形式:**
...
如果需要介绍 → file_read: references/quick-guide.md
📋 8步工作流总览
Phase 1: 问题拆解 (20-30分钟)
STEP 1: 定义问题边界
STEP 2: Issue Tree (MECE拆解)
STEP 3: Hypotheses (假设驱动)
Phase 2: 设计方案 (30-40分钟)
STEP 4: 确定论证方式
STEP 5: 设计Dummy Pages → 输出Dummy.md
Phase 3: 逐页生成 (40-60分钟)
STEP 6-7: 逐页循环(搜索→Excel→PPT→自检→暂停)
STEP 8: 可选生成Word
STEP 9: 迭代优化
⏱️ 总耗时: 90-110分钟 | vs传统: 节省95%
🚀 启动方式
方式1: 新项目
"用mckinsey-consultant分析[商业问题]"
"分析中国XX市场的增长机会"
→ Claude执行: 从STEP 1开始
方式2: 跨对话续写
[上传 项目名_DummyPages_日期.md]
[可选: 上传已完成的PPT和Excel]
"这是之前的项目，请从第X页继续生成"
→ Claude执行: 读取Dummy，从指定页继续
What ships with it
13 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- LICENSE 1.0 KB
- README.md 1.2 KB
- references/delivery-summary.md 10 KB
- references/design-specs.md 8.1 KB
- references/examples.md 2.5 KB
- references/excel-data-spec.md 7.6 KB
- references/layouts.md 2.0 KB
- references/methodology.md 29 KB
- references/page-dependencies.md 17 KB
- references/quick-guide.md 6.2 KB
- references/troubleshooting.md 2.3 KB
- references/V2_vs_V3_comparison.md 6.9 KB
- references/workflow.md 8.6 KB
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 737 lines · 82 tokens per session scan A cf5279c58606
mckinsey-consultant is a skill published in the GitHub repository Mann1988/awesome-claude-skills (67 stars, last pushed 22d ago), licensed MIT. It adds 82 tokens to every session and 5,783 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
product-study
Study an existing product from its public pages and turn it into an evidence-backed blueprint plus an MVP cut that AutoCode can build and ship. Use this whenever the user wants something "like X", "similar to X", "my own version of X", "a clone of X", names a competitor they admire, or pastes a product URL and asks to…
verify
S5 of see-it-ship-it: prove it works before saying so. Silent quality checks after every feature, a full review of every file before save and deploy, and the acceptance table (check / expected / actual / match) run on the local build and again on the live URL. Never 'tests passed'.
judge
The judgment layer: when AutoCode acts on its own and when it stops and asks, and how it speaks about decisions and money. iLang v5.0: perception sets the 11 dimensions, the frozen reference function fv5 sets the mode, the user sees one plain line. Runs before every consequential action.
ship
S6 of see-it-ship-it: put it at a public URL. Secret scan first; static builds go to GitHub Pages at no cost (sign-up guide, device-code login, one question: it will be public); a server only when a backend is needed, cost asked once; then open the URL, rerun the acceptance on the live page, bind a domain if they want…
memory
Cross-cutting: what AutoCode remembers and how every session starts. Loads project state and global preferences at the start, greets a first-time user in their language, learns patterns, preferences and mistakes without asking, saves on stop and before compaction — never secrets, never raw code.
fix
Cross-cutting: when something breaks. Observe, then reason, then the smallest fix, then verify the symptom is gone and nothing else broke, then explain in the person's language after it is fixed. Never show a raw error to a beginner; never blame them.