Why this repo exists

Most method libraries stop at definitions. That is not enough for an AI agent. The agent also needs to know when a method fits, what input is missing, how to work through the steps, and what to produce at the end.

Otherwise, familiar problems show up:

What goes wrong What the skill adds
The agent reaches for SWOT whenever a question sounds strategic Triggers and When Not To Use rules narrow the choice
Facts, opinions, and assumptions get mixed together Evidence requirements expose confidence and gaps
The analysis sounds polished but changes nothing The output has to support an actual decision
Several frameworks are stacked for show A method stays only if it changes the action, risk, validation, or deliverable
The answer sounds certain even though key inputs are missing Quality gates make assumptions and validation actions visible

Each skill is a standalone working guide. It tells an agent how to run the method, not just how to describe it.

30-second start

Install the library with the open Skills CLI:

npx skills add fzfclee/consulting-skills

Or install one method with GitHub CLI:

gh skill install fzfclee/consulting-skills systems-thinking --agent codex --scope user

Then ask your agent:

Use systems-thinking to diagnose why this issue keeps returning after repeated fixes.
Separate evidence from assumptions. Identify feedback loops, delays, leverage points,
side effects, and the next validation signal.

Every skill follows the same basic pattern:

Business question
      ↓
Check inputs and use boundary
      ↓
Execute explicit method steps
      ↓
Produce a reusable decision artifact
      ↓
Run the method-specific quality gate

What can you use these skills for?

You do not need to be a consultant or memorize a framework. Give your agent the situation, the facts you have, what you are unsure about, and the decision you need to make. Then start with one relevant skill.

A question you might have Useful starting skills What you get back
Should I stay in this job, take an offer, or keep looking? evidence-map, scenario-planning Facts and assumptions, plausible futures, and conditions for changing course
Which product, supplier, or large purchase should I choose? decision-matrix, cost-benefit-analysis A comparison that shows tradeoffs, weak evidence, and hard constraints
Why does the same problem keep coming back at work? change-event-timeline, systems-thinking A timeline, recurring pattern, feedback loops, and a safer intervention
Does a side business or new product make economic sense? business-model-canvas, break-even-analysis Key assumptions, the break-even threshold, and what to test before spending more
Why are people not using a product or internal tool? jobs-to-be-done, metrics-tree Adoption hypotheses, missing evidence, actions, and success measures
How do I get a proposal through a complicated group? stakeholder-power-map, communications-plan Decision makers, informal influence, engagement moves, and a communication plan

Use one method first. Add a second only when it answers a different question that could change the decision.

Choose by situation

Start with the problem you have, not the framework you remember.

Situation Start with What it should produce
Facts, opinions, and assumptions are mixed evidence-map Evidence ledger and confidence gaps
A broad problem needs a clean structure issue-tree Decision-oriented problem tree
A problem keeps returning systems-thinking Loops, delays, leverage points, side effects
Several causes remain plausible abductive-reasoning Competing explanations and discriminating tests
Multiple options need comparison decision-matrix Criteria, tradeoffs, sensitivity, recommendation
Stakeholders hold hidden veto power stakeholder-power-map Influence map and engagement moves
Product adoption or realized value is weak jobs-to-be-done Demand-side diagnosis and adoption hypotheses
A plan needs a failure rehearsal pre-mortem Failure modes, prevention, triggers, contingencies
The future depends on critical unknowns scenario-planning Robust moves, contingent moves, signposts
Work is blocked by one limiting factor constraint-analysis Constraint proof and exploitation plan

If the right starting point is still unclear, browse catalog.yaml or look at the seven controlled comparisons.

The 58-skill library

Method family Skills Typical decisions
Problem framing and evidence 8 What is true, missing, noisy, or poorly framed?
Reasoning and root cause 8 Which explanation is strongest, and why?
Systems, risk, and futures 5 What could persist, fail, or change over time?
Strategy, market, and commercial 10 Where to play, how to compete, and how to win?
Customer, product, and experience 7 What do users need, adopt, value, or reject?
Prioritization and economics 7 Which option deserves time, money, or sequence priority?
Measurement and performance 3 What should be measured and governed?
Stakeholder, change, and governance 7 Who decides, influences, owns, or resists?
Execution and validation 3 What is the smallest credible action and proof?

Problem framing and evidence

5w1h-analysis · affinity-diagram · assumption-inventory · evidence-map · issue-tree · mece-framework · mind-map-analysis · signal-vs-noise-filter

Reasoning and root cause

abductive-reasoning · constraint-analysis · deductive-reasoning · first-principles-thinking · fishbone-diagram · five-whys-root-cause · hypothesis-tree · inductive-reasoning

Systems, risk, and futures

critical-uncertainties · pre-mortem · risk-matrix · scenario-planning · systems-thinking

Strategy, market, and commercial

account-plan · business-model-canvas · competitive-positioning · deal-strategy-map · go-to-market-diagnosis · pestel-analysis · porter-five-forces · pricing-strategy-check · swot-analysis · win-loss-review

Customer, product, and experience

customer-segmentation · customer-success-health-score · empathy-map · jobs-to-be-done · kano-model · service-blueprint · user-journey-mapping

Prioritization and economics

break-even-analysis · cost-benefit-analysis · decision-matrix · effort-impact-matrix · rice-scoring · weighted-scorecard · wsjf-prioritization

Measurement and performance

balanced-scorecard · metrics-tree · north-star-metric

Stakeholder, change, and governance

change-event-timeline · change-impact-analysis · communications-plan · force-field-analysis · power-interest-grid · raci-matrix · stakeholder-power-map

Execution and validation

fifteen-percent-solutions · min-specs · validation-plan

How quality is checked

Every skill must include five practical sections:

  1. Required Inputs
  2. When Not To Use
  3. Step-by-Step Execution
  4. Output Template
  5. Quality Gate

The repository checks those sections automatically:

Check Current coverage
Standalone skills validated 58
Evaluation prompts covering all skills 70
Method-selection discrimination cases 24
Controlled comparisons 7
Neutral catalog entries 58
Portability or local-path failures allowed 0

The validator checks names, frontmatter, required sections, links, UTF-8 text, catalog consistency, and portability. These are structural checks. They catch broken packages and weak method contracts, but they do not prove that a model will make the right decision. The evaluation set therefore keeps its prompts and scoring rubric in the open.

Seven controlled comparisons

Decision problem Example route
Clarify an ambiguous problem Evidence Map → Issue Tree
Diagnose a recurring issue Timeline → Systems Thinking → Constraint Analysis
Prioritize competing options Decision Matrix → Risk Matrix
Build a stakeholder strategy Stakeholder Power Map → Account Plan
Improve product adoption and ROI Evidence Map → JTBD → Metrics Tree
Assess market entry PESTEL → Five Forces → Positioning
Evaluate a career decision No method vs Evidence Map → Stakeholder Power Map → Scenario Planning → Weighted Scorecard

These examples are not a universal router or a scientific model benchmark. Each one runs the same input twice: first as a direct AI answer with no named method, then with the listed method chain. The comparison keeps the evidence boundary, method workpapers, handoff between steps, final decision artifact, action plan, success measures, and reversal conditions.

What does a method chain add?

In the seven comparisons, the direct answer often reaches a sensible recommendation quickly. The method-based run may reach the same conclusion. The difference is that it leaves behind workpapers, the handoff from one method to the next, owners and timing, success measures, decision gates, and triggers for changing course.

A method chain makes the reasoning easier to review and the plan easier to run. It does not make the conclusion automatically correct.

What this is, and what it is not

This repository is This repository is not
A portable library of standalone consulting method skills A claim of ownership over established methods
A set of execution instructions for AI agents A substitute for evidence or professional judgment
A neutral catalog with explicit boundaries A rule that every problem needs a framework
An open method resource in the O2V Framework ecosystem The complete O2V Framework or CLEAR / Signal-to-Action methodology

Consulting Skills is initiated by Li Zhi and maintained as an open method library in the broader O2V Framework knowledge ecosystem. O2V Framework and CLEAR / Signal-to-Action are maintained separately as complete methodology assets.

Repository structure

skills/<skill-name>/SKILL.md   58 standalone method skills
catalog.yaml                   neutral discovery metadata
examples/                      seven controlled decision comparisons
evaluations/                   behavior and method-selection test cases
scripts/                       catalog generation and validation

Contributing

Contributions should improve method fidelity, trigger clarity, evidence discipline, output usability, or decision impact. A new method should fill a real decision need rather than increase the count.

Read CONTRIBUTING.md, open an issue, or submit a pull request.

Intellectual property

This project does not claim ownership of established consulting tools, analytical methods, academic theories, or management frameworks. Method names are used descriptively; their rights and academic contributions remain with their respective creators, researchers, institutions, and rights holders.

Apache License 2.0 applies to this repository's original skill text, executable structure, examples, catalog, and validation code. O2V Framework, CLEAR, Signal-to-Action, and their original framework expressions and assets are separate works by Li Zhi and are not licensed under this repository's Apache-2.0 license.

See NOTICE and ATTRIBUTIONS.md.

Contact

Official website: www.o2vframework.com

Email: [email protected]

Found it useful? A star helps others find it. Found a weak method? Open an issue.