the skill
Consulting Engine
Router for the twelve consulting-engine method skills. Read this first when an engagement-level request could fit several skills, when the user asks "which skill should I use for this", "where does this sit in the engine", "what comes next after X", or when the question mentions the engagement lifecycle (frame, evidence, choose, execute, measure, communicate, sell). Also read to settle the boundary with product-strategy and strategy-toolkit. Holds the engagement map, the three "what would have to be true" objects, and the hand-off contracts once, so the method skills do not each carry a copy.
Twelve method skills, one per step of a strategy engagement. Each does one thing and hands a named artefact to the next. This file is the map; the method is in each skill.
Engagement map
External skills the engine calls: boardroom-storm-research for multi-perspective desk research; pptx, docx, xlsx for files. A pre-mortem on a chosen plan is not in the engine; run it separately and carry back only the top mitigations.
Routing rules
-
Object of "what would have to be true". A day-one hypothesis before evidence:
problem-framing(assumptions flagged, not tested). An explanation of what is true now:hypothesis-stress-testing(what observation kills it). A strategic option:strategic-choice(owns the phrase). A chosen plan: pre-mortem, not the engine. -
"How could we be wrong." If the sentence contains "the plan", "the launch", "should we go ahead": pre-mortem. "The cause", "our thesis", "is it true that":
hypothesis-stress-testing. "What will competitors do", "which future":war-gaming. -
Evidence by source. Numbers in accounts:
value-driver-modeling. Industry structure and rivals:market-competitive-intelligence. Supplied documents:document-synthesis. Interviews and observation:primary-research-synthesis. Sources to be found:boardroom-storm-research. -
Sequence. No problem statement: frame first. No options: choose before gaming or executing. No charters: execute before measuring. Analysis not done: do not communicate yet.
-
product-strategy boundary.
product-strategyowns the answer to a product or proposition question (where to play, what to build, why it wins, how it makes money, how to launch) and its deck architecture. It calls engine skills as tools:problem-framingfor the issue tree,market-competitive-intelligencefor Phase 1 fact base,value-driver-modelingfor the business case model,strategic-choicefor testing proposition options,kpi-architecturefor definition sheets and the ledger,executive-communicationfor the pyramid and titles. Enterprise-wide, non-product questions stay in the engine. -
strategy-toolkit boundary.
strategy-toolkitis the breadth layer: 21 short moves that give a structured answer in one session, with no gates, templates or worked examples. Route there when the user wants a quick structured take, one named move (a profit-pool read, a decision memo, a stakeholder map) or a McKinsey, Bain or BCG re-flavoring. Route here when the work needs a gated method or a hand-off artefact. Each toolkit move names its engine counterpart under "Long form".
Conventions every engine skill follows
- Deliverable first, headline in the first 150 words; assumptions beside the content they support, never as an opening list.
- One round of clarifying questions when a load-bearing input (decision-maker, target, deadline, data) is missing; otherwise proceed on labelled assumptions.
- No method vocabulary in what the client reads: no stage numbers, checkpoint labels, framework authors, skill names, hand-off notes, or the worked example's client.
- Each artefact appears once; tables carry the value, prose around them is trimmed first.
- Quality bar run silently; working notes only on request or when they change a decision, after the deliverable under one heading.
- Hand-offs are one sentence at the end in the user's terms, not a table in the pack.