Enterprise AI Skills

13 consulting frameworks from a former McKinsey consultant, packaged as drop-in Claude skills. Issue trees, decision memos, McKinsey-style charts, full deck pipelines — the stuff strategists, PMs, and ops leaders run anyway, now executable in one prompt.

A 30-second walkthrough: take a real personal question ("Which sport should I pick up?"), run it top-down through 4 of the skills — Issue Tree → Storyline → Charts → Deck Pipeline — and out comes an 8-slide deck with native editable charts.

>

Full worked example, including the prompts: examples/which-sport-to-pick-up/.

📖 Prefer to scan at your own pace? Open the storyboard (no autoplay).

Latest update (May 2026): added Synthesis, Hypothesis Tree, Stakeholder Map, Top-Down Memo, and Workshop Designer; replaced the Gamma-based deck path with native python-pptx + editable charts.

---

Try it in 30 seconds

Open Claude (or any LLM with these skills loaded) and paste:

Build me a McKinsey-style deck on a decision I'm weighing:
"Which sport should I pick up: pickleball, tennis, or running?"
Run the full top-down workflow: 
Use issue-tree-builder to decompose into MECE branches
 
Use storyline-builder to turn it into 8 claim-shaped slide titles
 
Use mckinsey-charts for the data slides (bar+callout, waterfall)
 
Use deck-pipeline (strategist → builder → critic → fixer) to produce the final .pptx

You'll get the same arc as the GIF above, with a real .pptx you can open and edit. Swap the question for one you actually care about: "Which side project should I commit to?", "Should I take the job?", "Where should we move?" — same workflow, your decision.

What one of the data slides looks like:

---

Install

• Claude Code: clone this repo — skills auto-load when you work in the directory.
• Claude.ai: download the .skill file for any skill → Settings → Skills → Add Skill.
• Any LLM: copy the contents of SKILL.md into your conversation as context.

git clone https://github.com/sruthir28/enterprise-ai-skills.git

No dependencies beyond Claude itself. The deck pipeline (deck-pipeline + mckinsey-charts) requires Python 3.9+ and python-pptx if you want to generate .pptx files locally — see examples/which-sport-to-pick-up/build.py.

---

Skills

Thinking

The McKinsey muscle — frame the problem, commit to an answer, pull the "so what" out of messy inputs.

| Skill | What It Does | Best For | |-------|-------------|----------| | SCPR Framework | Structure arguments using Situation-Complication-Problem-Recommendation | Exec summaries, memos, strategic recs | | Issue Tree Builder | Break down problems into MECE components | Analysis, case prep, work plans | | Hypothesis Tree | Commit to a Day-1 answer and design tests that could kill it | Start of any analysis where you'd otherwise "boil the ocean" | | Synthesis | Turn a pile of interviews / transcripts / research into 3 insights with evidence + so-what | After customer interviews, surveys, research dumps | | Prioritization | Score features using RICE, Impact/Effort, Value/Complexity | Roadmap decisions, feature trade-offs | | AI Use-Case Scorer | Score AI use cases on Value × Feasibility × Safety; tier them Do Now / Quarter / Park / Avoid | Personal AI roadmap, "what should I AI-ify next" | | Data Insights | Analyze any dataset and answer key questions in plain English | Quick analysis without being a data scientist |

Writing

Answer-first, evidence-dense, decision-forcing.

| Skill | What It Does | Best For | |-------|-------------|----------| | Decision Memo Builder | 1-page memo that forces a yes/no (context → options → rec → risks → ask) | Leadership decisions, scope cuts, buy/build calls | | Top-Down Memo | Answer-first writing using the Minto Pyramid — lead, 3 arguments, evidence | Emails to execs, briefing docs, Slack messages that need to land | | Storyline Builder | Build slide narratives where each line = a slide title | Decks, pitch presentations, board updates |

Meetings & People

Where the politics decide the outcome.

| Skill | What It Does | Best For | |-------|-------------|----------| | Meeting Prep Kit | Pre-read + agenda + talking points + objections for any meeting you're driving | Exec reviews, stakeholder pitches, scope negotiations | | Stakeholder Map | Map the room onto Power/Interest with each person's stance + one move | Cross-functional decisions, reorgs, launches | | Workshop Designer | Design a half-day or full-day session that ends with a committed artifact + owners | Offsites, leadership alignment, quarterly planning |

Decks

Native, editable PowerPoint — no Gamma, no images.

| Skill | What It Does | Best For | |-------|-------------|----------| | McKinsey Charts | Three workhorse charts (bar+callout, stacked-over-time, waterfall) as native python-pptx objects | Any chart that needs to look EM-reviewed, not Excel-default | | Deck Pipeline | 4-agent pipeline: Strategist → Builder → Critic → Fixer. Uses Storyline + Charts to produce a critique-proof deck. | Board decks, strategy presentations, any high-stakes slides |

Agents

| Skill | What It Does | Best For | |-------|-------------|----------| | McKinsey Critic | Reviews decks and docs like a McKinsey engagement manager — grades, flags problems, gives top 3 fixes | Quality gate before anything goes to stakeholders |

---

How the skills compose

Most real work uses 2–3 of these together. Some common stacks:

• Customer research → memo: Synthesis → Top-Down Memo (or Decision Memo) → McKinsey Critic
• Strategic question → recommendation: Issue Tree → Hypothesis Tree → run the tests → Decision Memo
• Big decision with politics: Stakeholder Map → Decision Memo → Meeting Prep Kit (for each 1:1)
• Board deck (the GIF above): Issue Tree → Storyline Builder → Charts → Deck Pipeline → McKinsey Critic
• Offsite or alignment workshop: Workshop Designer → Meeting Prep Kit (pre-1:1s) → workshop runs → Top-Down Memo (post-send)

---

What this is, and isn't

Is: a library of 13 prompts-with-structure that turn a fuzzy strategic ask into a decision-shaped artifact. Each skill encodes a specific consulting move (MECE decomposition, Day-1 hypothesis, pyramid principle, etc.) so Claude doesn't reinvent the structure every time.

Isn't: a chatbot, a UI, or generic "AI for work" prompts. There's no model fine-tuning. The skills work because the structure is right, not because the model is special.

---

Need this adapted for your team?

Domain-specific versions, custom workflows, AI rollouts — reach out: sru281+build@gmail.com.

---

MIT licensed. Built by Sruthi Chintakunta.