MECE Issue Tree Skill for Claude Opus
A downloadable Claude skill that makes Opus 4.6 decompose any business problem into a clean MECE issue tree. Written in the McKinsey diagnostic discipline: no overlapping branches, no gaps, no guesswork.
Free skills and prompts for Claude and strategy work
Templates for Claude, ChatGPT and Perplexity — from diagnostics to board-ready decks.
What MECE discipline requires
MECE stands for Mutually Exclusive, Collectively Exhaustive. It is the structural discipline McKinsey, BCG, and Bain use to decompose any business problem before running analysis. Each branch of the issue tree must cover a distinct portion of the problem space (no overlap) and together all branches must cover the full space (no gaps). Violate either condition and the diagnosis will miss the root cause or double-count effort.
In practice, MECE issue tree discipline is hard to maintain under time pressure. The temptation is to list symptoms rather than decompose causes, collapse levels that belong at different depths, or merge branches that are actually separable. A well-built issue tree forces a team to resolve those tensions before the first hypothesis is written.
This skill encodes that discipline into Claude. It is part of the broader Claude skills for strategy collection, which covers diagnosis, execution, and stakeholder alignment.
What the skill does
The skill is a markdown file you add to a Claude Project as knowledge. Once loaded, it instructs Claude Opus 4.6 to run MECE decomposition on any business problem you describe. It introduces four specific behaviors that ad-hoc prompting does not reliably produce.
Two-level minimum
Forces a decomposition from problem to drivers to sub-drivers before stopping.
MECE test at each branch
Explicitly checks for overlap between sibling branches and tests whether together they cover the full solution space.
Diagnostic questions per branch
Each branch comes with the key question it answers and the data required to close it.
Assumption flagging
Surfaces untested beliefs embedded in the tree before the analysis begins.
The output is a structured issue tree you can paste directly into a brief, a Word document, or feed into Oria to render as a diagnostic slide.
Install the skill in Claude
Download the skill file
The skill is a single .md file. Download it from the claude-skills-for-mckinsey-analysis repository. It lives in the McKinsey-skills-diagnosis folder alongside the other diagnostic skills.
Create a Claude Project
Go to claude.ai, open the left sidebar, click Projects, then Create Project. Name it something like "Strategy Diagnosis" or "Issue Tree Assistant."
Add the skill as Project Knowledge
Inside the project, open Project Knowledge, click Add Content, and upload the .md file. Claude reads it automatically in every conversation inside that project. You can add additional diagnosis or execution skills from the same repo at the same time.
Start a conversation inside the project
Open a new conversation inside the project, describe the business problem, and ask Claude to run the issue tree. No copy-paste of the prompt file required. Claude sees it from project knowledge.
How to use it
Describe the problem clearly. Provide the business context, scope, and any constraints you already know. Then ask Claude to build the issue tree explicitly referencing the skill. A prompt that works consistently:
Example prompt
"Build a MECE issue tree for the following problem using the issue tree skill in the project. Decompose to at least two levels. Apply the MECE test at each branch and flag any data gaps or untested assumptions before closing the tree. Problem: [describe your business problem here]."
Tip
Once Claude returns the tree, you can expand specific branches by asking it to decompose a single node further, or redirect by marking a branch as "resolved" and asking it to focus on the others. The skill handles partial updates without restarting the full tree.
This skill pairs well with the consulting-grade problem solving skill, which handles the upstream framing before the tree is built.
When to run it
Run this skill at the start of any diagnostic workstream, before writing a hypothesis or assigning analysis. It is most useful when:
When the tree is done, the next step is usually a slide. Issue trees map directly to hierarchical diagram layouts, MECE grid structures, and multi-branch diagnostic visuals. These are complex layouts where Oria's Visual Rendering engine produces fully editable PowerPoint output rather than the fragile HTML renders typical HTML-based generators produce. Paste the tree into Oria's Text to Slide, describe the layout, and Oria builds the slide inside PowerPoint in under three minutes.