Thinking Framework Skills

An evidence-graded library of agent-executable thinking-method skills.

Every method is reduced to its working mechanism, graded honestly on how strong its evidence actually is, and shipped as a skill that produces a concrete artifact, not prose.

What it is  ·  Install  ·  Frameworks  ·  Evidence  ·  Recipes  ·  Live site

🧠 What this is

AI agents are fluent and fast, and surprisingly weak at the moves that make thinking actually good: reframing a problem before solving the wrong one, separating evidence from inference, imagining how a plan fails before it does, stress-testing a decision from more than one angle. Humans are not much better under time pressure. Both converge too early.

thinking-framework-skills packages the durable core of the structured-thinking tradition (decision science, creativity research, systems thinking, foresight, critical thinking) as small, composable, agent-ready skills. Each one helps a person or an agent reframe a problem, generate options, challenge an assumption, trace a consequence, or stress-test a decision, and hands back a usable artifact.

Three things make it different from a list of mental models:

It is It is not
Mechanism-first - the durable cognitive move, named for what it does A museum of trademarked frameworks
Evidence-graded - an honest tier (S/M/P/V/A/C/X) on every skill, including "weaker than people think" A confident claim that every method is "proven"
Artifact-producing - a risk register, an option matrix, an argument map, a Thinking Plan A set of vibes-y prompts
Composable - skills chain into recipes, passing a compressed artifact at each step A pile of unrelated one-offs
Honest about misuse - every skill names where it misleads ("when NOT to use this") A cargo-cult checklist

Relationship to pm-skills: sibling library, no technical coupling. thinking-framework-skills helps decide what to work on and why it is sound; pm-skills helps execute how. They compose; neither depends on the other.

⚡ Quick start

Claude Code (recommended):

/plugin marketplace add product-on-purpose/agent-plugins
/plugin install thinking-framework-skills@product-on-purpose

All 63 frameworks (plus the 4 tools and 9 recipes) become available immediately, invocable by name (for example /think-premortem).

Cross-agent (Cursor, Copilot, Cline, and others via the open skills CLI):

npx skills add product-on-purpose/thinking-framework-skills

Clone or download:

git clone https://github.com/product-on-purpose/thinking-framework-skills.git

Your first run. Pick a real decision you are about to commit to, then:

/think-premortem "we're about to launch a free tier to drive signups"

You get a ranked risk register: for each top risk, a leading signal, a mitigation, an owner, and a kill criterion. That artifact, not a feeling of caution, is the point. You do not need an agent - every skill is a procedure you can run by hand with the template on its page.

Not sure which framework you need? Start with the Framework Advisor: describe your situation in plain language and it returns a prioritized Thinking Plan of which skills to run, in order, and what to skip. It is the front door to everything else.

How the Framework Advisor fits into your workflow:

%%{init: {'theme':'base','themeVariables':{'primaryColor':'#eef2ff','primaryBorderColor':'#c7d2fe','lineColor':'#6366f1'}}}%%
flowchart LR
  sit["Your situation<br/>(a prompt)"]:::in --> adv{"Framework Advisor<br/>(the front door)"}:::router
  adv --> plan["Thinking Plan<br/>(prioritized, evidence-graded)"]:::plan
  plan --> s1["run skill 1"]:::skill
  plan --> s2["run skill 2"]:::skill
  s1 --> art["Artifact(s):<br/>risk register, option matrix,<br/>argument map"]:::art
  s2 --> art
  sit -. "or apply directly" .-> top3["think-top3 /<br/>think-random-frameworks"]:::skill
  top3 --> art
  classDef in fill:#dcfce7,stroke:#86efac,color:#166534;
  classDef router fill:#ddd6fe,stroke:#a78bfa,color:#4c1d95;
  classDef plan fill:#fef9c3,stroke:#fde047,color:#854d0e;
  classDef skill fill:#eef2ff,stroke:#c7d2fe,color:#3730a3;
  classDef art fill:#e0f2fe,stroke:#7dd3fc,color:#075985;

📖 Full walkthrough: docs/getting-started.md · Explore the whole library: the live site.

🔬 What makes it different

The field of "thinking methods" has three uneven layers: a small empirical core with replicated study evidence, a large practitioner ring of long-standing heuristics with real traction but limited formal validation, and a noisy outer ring where popularity and weak evidence get conflated. Most libraries flatten all three into one shiny catalog.

This one does the opposite, and that honesty is the product:

  • Grade evidence transparently. Every skill carries a tier, and its dossier states what the research does and does not show. A practitioner-tier method labeled honestly is more trustworthy than one dressed up as science.
  • Do not launder statistics. The often-cited "premortems surface ~30% more reasons" measures the number of reasons, not a 30% gain in decision quality. Claims are scoped to what the studies actually measured.
  • Mechanism over ritual. The skill implements the durable move and names the branded ritual as lineage, never the reverse. (So the library ships Parallel Perspectives Review, not the trademarked Six Thinking Hats.)
  • Flag transferred evidence. Almost no studies test an AI agent running these methods. Where evidence comes from human-subject research, the page says so.

🗺️ The library at a glance

63 frameworks across 12 cognitive-operation families (56 core plus 7 contested lenses), arranged as a thinking lifecycle. (Four tools and nine recipes ride on top.) You rarely run all twelve; the Framework Advisor picks the few that fit your situation.

%%{init: {'theme':'base','themeVariables':{'primaryColor':'#eef2ff','primaryBorderColor':'#c7d2fe','lineColor':'#6366f1','fontFamily':'system-ui, sans-serif'}}}%%
flowchart TB
    classDef frame fill:#e0e7ff,stroke:#6366f1,color:#1e1b4b,font-weight:bold
    classDef gen fill:#dcfce7,stroke:#16a34a,color:#14532d,font-weight:bold
    classDef analyze fill:#fef9c3,stroke:#ca8a04,color:#713f12,font-weight:bold
    classDef decide fill:#ffedd5,stroke:#ea580c,color:#7c2d12,font-weight:bold
    classDef reflect fill:#f3e8ff,stroke:#9333ea,color:#581c87,font-weight:bold

    PF["1. Problem Framing (6) - frame the real problem"]:::frame
    DI["2. Divergent Ideation (6) - generate options"]:::gen
    PM["3. Perspective & Multi-Lens (2) - see it from other angles"]:::gen
    SC["4. Systems & Consequences (9) - trace consequences"]:::analyze
    AB["5. Assumption & Belief Challenge (5) - challenge assumptions"]:::analyze
    RC["6. Reasoning Clarity (5) - clarify the reasoning"]:::analyze
    DO["7. Decision & Option Evaluation (12) - decide between options"]:::decide
    SO["8. Strategy & Opportunity (2) - test strategy against alternative futures"]:::decide
    EV["9. Ethics & Values Deliberation (4) - weigh the moral trade-off"]:::decide
    RR["10. Risk & Resilience (4) - anticipate what could go wrong"]:::decide
    SY["11. Synthesis (4) - turn inputs into a message"]:::reflect
    MR["12. Meta-Thinking & Reflection (4) - learn and route"]:::reflect

    PF --> DI --> PM --> SC --> AB --> RC --> DO --> SO --> EV --> RR --> SY --> MR

In text: frame the problem, generate options, see it from other angles, trace consequences, challenge assumptions, clarify the reasoning, decide between options, test strategy against alternative futures, weigh the moral trade-off, anticipate what could go wrong, synthesize, then reflect. See the full color-coded map (by evidence tier) on the live site.

🔬 The evidence model

Honest grading is the differentiator, so the key comes before the catalog: every skill and every claim is labeled with one of seven tiers, from strongest to weakest.

Tier Meaning
S Strong - replicated experimental or meta-analytic support
M Moderate - real evidence, but narrower, correlational, or field-based
P Practitioner - widely used and defensible, without strong controlled evidence
V Vendor - originates from a consultancy or branded methodology
A Anecdotal - case reports and testimonials
C Conceptual - reasonable, not yet demonstrated
X Poor/contradictory - the evidence cuts against it (excluded, documented)

A few skills carry a split grade (for example M/P or S/M): the mechanism rests on one tier while a specific claim about it rests on another. Where a grade leans on human-subject research that has not been tested on an AI agent, the skill's dossier flags that transfer explicitly rather than overclaiming.

A strong-evidence core anchors the library; everything else is honestly labeled around it. The bibliography aggregates the graded sources so a skeptic can trace any claim to its grounding. See docs/concepts.md for the short version.

📚 The catalog

All 63 frameworks, by family. Seven are contested lenses (SWOT, ACH, Five Whys, Eisenhower/MoSCoW/Pareto, Complexity Domain Sort, Reflective Equilibrium, QCA): famous-but-weak methods graded honestly and shipped caveat-first, explicit-request-only, marked Contested lens in their rows. The Tier column is the evidence grade defined just above. Each name links to its full page - mechanism, numbered procedure, worked example, and graded sources - on the live site. (The routers and applicators are listed under Tools; the chains under Recipes.)

Problem Framing - frame the real problem (6)

Skill Tier What it does
Problem Restatement M/P Rewrite the problem several ways to expose hidden framing, then pick a more useful one
Abstraction Laddering P Move up ("why") and down ("how") the ladder to find the altitude where the problem is workable
Contradiction Resolution M/P Reframe a trade-off as a contradiction to dissolve via separation in time / space / scale / condition, with an honest exit when it is genuinely real
Boundary Critique C/P Audit the boundary judgments behind a frame (who benefits, decides, counts, has standing) in is vs ought, naming who is affected but excluded
Frame Creation C/P Distil themes and the core paradox, then abduce a new "as if it were Y" standpoint and reason forward to solutions
Five Whys X Contested lens (caveat-first, explicit-request-only). Reliable only for simple linear single-cause failures and misleading beyond them (Card 2017); leads with that caveat, flags where a problem is multi-cause, and points multi-cause analysis to issue-tree

Divergent Ideation - generate options (6)

Skill Tier What it does
Morphological Analysis P Lay out a solution's independent parameters and their possible values as a Zwicky box, then cross-combine and prune to internally consistent configurations
Brainwriting S Silent, parallel, written idea generation that reliably outperforms verbal brainstorming
Far-Analogy Ideation S Transfer solutions from distant domains, which produce more original ideas than near ones
SCAMPER P Run an idea through seven transformation prompts to force structured variation
Question Burst P Generate a rapid burst of questions, rank them, and pursue the most catalytic one
Assumption Reversal P Surface the assumptions baked into a problem, negate them, and generate non-obvious reframes

Perspective & Multi-Lens - see it from other angles (2)

Skill Tier What it does
Role-Storming P Generate ideas while inhabiting a chosen non-self persona, using the assumed identity to lower self-censorship and shift associations
Parallel Perspectives Review P Examine a decision through several separated lenses in turn, then synthesize a balanced read

Systems & Consequences - trace consequences (9)

Skill Tier What it does
Three Horizons C Hold three time-horizon curves at once - declining present, contested transition, emerging future - and locate the actor in the transition zone
Process Tracing P Test rival causal explanations of a single case by each piece of evidence's diagnosticity (hoop, smoking-gun, straw-in-the-wind, doubly-decisive tests)
Causal Layered Analysis C Read an issue down four layers (litany, system, worldview, myth) and reconstruct a preferred future back up each, anchored by a deliberately changed metaphor
Stocks and Flows Reasoning S Reason explicitly about accumulations and rates, which people systematically misjudge
Causal Loop Diagrams M/P Close and sign the feedback loops (reinforcing or balancing) to read why a system spirals, settles, or oscillates
Futures Wheel P Map first-, second-, and third-order consequences radiating from a change
Iceberg Model P Move from events down to the patterns, structures, and mental models that produce them
Theory of Constraints P Find the single binding constraint capping throughput and attach its exploit / subordinate / elevate decisions (the five focusing steps)
Qualitative Comparative Analysis (QCA) P Contested lens (warn-and-redirect, explicit-request-only). Certifies configurations from noise at session scale (Lucas and Szatrowski 2014); does not build the truth table, it warns and routes to reference-class-forecasting

Assumption & Belief Challenge - challenge assumptions (5)

Skill Tier What it does
Authentic Dissent S Cultivate genuine minority disagreement, which improves reasoning where role-played dissent does not
Consider the Unknowns M Before committing to a judgment, explicitly list the relevant variables you cannot observe and weigh the gap they leave
Ladder of Inference Check P Trace how you climbed from raw data to conclusion to catch where interpretation crept in
Red Team Light P A lightweight adversarial pass that attacks a plan to surface its weak points
Analysis of Competing Hypotheses (ACH) X Contested lens (warn-and-redirect, explicit-request-only). Controlled trials found ACH raises confidence without accuracy, so it does not build the matrix as if valid; it leads with that evidence and routes to a better-grounded move

Reasoning Clarity - clarify the reasoning (5)

Skill Tier What it does
Argument Mapping S Diagram the structure of claims, reasons, and objections to expose where it is weak
Argumentation Schemes P Identify which stereotyped argument pattern is in play, then test it with that scheme's standard critical questions
Natural-Frequency Bayesian Framing S Express probabilities as natural frequencies (3 in 1,000) to make conditional reasoning tractable
Evidence vs Inference Sort P Separate what is actually known from what is being inferred, and label each
Issue Tree P Decompose a question into a logical tree of sub-questions to make analysis tractable

Decision & Option Evaluation - decide between options (12)

Skill Tier What it does
Pairwise Comparison P Rank options with no absolute scale by judging every pair head-to-head, deriving the order from the win-counts with a consistency check
Dialectical Bootstrapping M Make an estimate, assume it is wrong and list why, estimate again from those changed assumptions, then average the two numbers
Interest-Based Negotiation P Separate interests from positions, decide against your BATNA inside the mapped ZOPA, and invent options for mutual gain before dividing value
Minimax Regret P Choose under deep uncertainty with no probabilities by minimizing the worst-case regret across the states of nature
Linear-Model Aggregation S Score options on a simple weighted model that tends to beat holistic judgment
Fermi Estimation M/P Estimate an unknown by decomposing it into order-of-magnitude factors, then multiplying back to a number with a low/high band
What Would Have to Be True P Turn a claim into the specific conditions that must hold, then test them
Decision Option Review P Compare options against weighted criteria with explicit tradeoffs
One-Way vs Two-Way Door P Classify a decision by reversibility and match the deliberation cost to it
Expected-Value Decision Tree P Price the uncertainty: a tree of choice and chance nodes, rolled back to an expected value per option, with a what-flips-it note
Eisenhower / MoSCoW / Pareto P Contested lens (caveat-first, explicit-request-only). A bundle of three weakly-evidenced prioritization presets; runs the requested one and leads with the laundering caveat (Zhu, Yang and Hsee 2018 measures the urgency bias, not the matrix as a remedy)
Complexity Domain Sort C Contested lens (caveat-first, explicit-request-only). The descriptively-named Cynefin sort; no controlled validity evidence (PMC 2021) and cargo-cult risk; leads with that caveat plus the Cynefin trademark and names what to do per domain

Strategy & Opportunity - test strategy against alternative futures (2)

Skill Tier What it does
Scenario Planning P Construct a set of divergent, internally consistent external futures (2x2 critical-uncertainty axes), then robustness-test strategy across them
SWOT X Contested lens (caveat-first, explicit-request-only). Leads with SWOT's weak evidence (Hill and Westbrook 1997), then forces the discipline a bare grid lacks, namely evidence-tagged and prioritized factors plus a TOWS matching step into options

Ethics & Values Deliberation - weigh the moral trade-off (4)

Skill Tier What it does
Veil-of-Ignorance Reasoning M Decide a values trade-off as if you had an equal chance of being any affected party, then return to the actual decision and confront the gap
Ethical Matrix P Grid affected parties (including voiceless ones) against prima facie principles (wellbeing, autonomy, fairness) and read the cell-level pattern of trade-offs
Reflective Equilibrium C Contested lens (caveat-first, explicit-request-only). Philosophically central but empirically untested as a procedure, with no checkable termination test (Brandt 1979); runs it with an explicit revision ledger and leads with the failure modes
Speculative Harms & Anti-Goals A Assume the design succeeds at scale, narrate how it harms third parties and is exploited in bad faith, then convert each harm into an explicit anti-goal

Risk & Resilience - anticipate what could go wrong (4)

Skill Tier What it does
Premortem S/M Imagine the plan has already failed and work backward to causes, tripwires, and kill criteria
Reference Class Forecasting S Estimate from the track record of similar past projects, not inside-view optimism
WOOP S Wish, Outcome, Obstacle, Plan: mental contrasting plus implementation intentions
Backcasting P Start from a desired future state and work backward to the steps to reach it

Synthesis - turn inputs into a message (4)

Skill Tier What it does
Contradiction / Tension Mapping C Map an interdependent polarity as a both/and - two poles, their upsides and downsides, a greater purpose, warning signs, and action steps - to manage rather than resolve it
Concept Mapping M/P Build a labeled-relationship concept network so each link reads as an explicit proposition, surfacing gaps and missing links
Affinity Mapping P Cluster many raw notes into emergent themes from the bottom up
Pyramid Principle P Structure communication as a governing claim over grouped, ordered support

Meta-Thinking & Reflection - learn and route (4)

Skill Tier What it does
After Action Review S Structured review of expected vs actual, and what to change, to improve the next loop
Interval Calibration Check P Test a stated confidence interval against an equivalent bet, widen until indifferent, then score stated confidence against the actual hit rate
Decision Journal P Record the decision, rationale, and prediction now to calibrate your judgment later
Belief-Update Routine P Re-score a standing inventory of open beliefs against new evidence on a cadence, with an explicit confidence delta and an under-updating guard

Browse them five other ways - by job, by evidence, by artifact, by situation, or on the map - in the site's Explore section. The skills themselves live in skills/.

🛠️ Tools (meta-skills)

Four skills are tools, not thinking methods: they operate over the library - routing you to the right framework, applying several at once, or researching new ones. They carry no evidence tier of their own (any grade in a tool's dossier is about the tool's behavior, not a method), and they live under /tools/ on the site.

Tool What it does
Framework Advisor Router, the front door. Describe a situation, get a prioritized Thinking Plan of which frameworks to run, in order, and what to skip. Use it when you do not know which method fits.
Top-3 Applicator. Rank the most relevant frameworks for your topic, apply the top three so each emits its artifact, then cross-synthesize. Use it when you want worked output now, not a plan.
Random Frameworks Applicator. Draw three frameworks at random and apply each regardless of fit, to break a frozen or fixated framing. Use it when the obvious lenses are not working.
Research Framework Engine. Research a candidate method, grade its evidence conservatively, and propose a catalog entry for review. This is how the library grows; it never auto-admits.

Frameworks, tools, and recipes, in one line: a framework is a single graded thinking method (the 63 in the catalog above); a tool helps you choose or apply frameworks (the 4 here); a recipe is a fixed chain of frameworks for a recurring job (the 9 below).

⚙️ How a skill works

Each skill is a self-contained unit: a portable SKILL.md (the mechanism and procedure), an evidence/dossier.md (the graded sources and honest caveats), a references/EXAMPLE.md (a worked example that sets the quality bar), and a skill.meta.yml sidecar (governance, taxonomy, relationships).

%%{init: {'theme':'base','themeVariables':{'primaryColor':'#eef2ff','primaryBorderColor':'#c7d2fe','lineColor':'#6366f1','fontFamily':'system-ui, sans-serif'}}}%%
flowchart LR
    classDef you fill:#1e293b,stroke:#0f172a,color:#fff,font-weight:bold
    classDef skill fill:#4f46e5,stroke:#3730a3,color:#fff,font-weight:bold
    classDef art fill:#166534,stroke:#14532d,color:#fff,font-weight:bold

    U["You<br/>a messy situation"]:::you
    S["A think- skill<br/>mechanism + procedure"]:::skill
    A["A concrete artifact<br/>risk register, option matrix,<br/>argument map, Thinking Plan"]:::art

    U -- invoke --> S -- produces --> A

When you run /think-premortem "...", the agent loads the skill, follows its numbered procedure, mirrors the worked example, and produces the artifact. No prompt engineering required. The docs site is a generated view of these files; see docs/architecture.md for how the skills become the site.

🧩 Recipes

Recipes chain several skills into one end-to-end job, passing a compressed artifact at each handoff. Nine ship today:

Recipe What it does
Kepner-Tregoe Find a deviation's cause, choose among defined options, and de-risk the rollout - the rational-process bundle run as a chain of shipped moves
PDCA / A3 Root-cause a performance gap, choose and run a countermeasure, then review actual versus expected and standardize or iterate
Reframe a problem Restate the problem, sharpen the question, and check the framing before you build
Expand options Reframe, then generate genuinely new options before judging any
Stress-test a decision Surface what must be true, weigh options, calibrate reversibility, and premortem the plan
Audit reasoning Separate evidence from inference, map the argument, and pressure-test it
First principles Decompose a problem to its fundamentals, then strip the inherited assumptions to rebuild from what is necessary
Idea-quality audit Score a batch of ideas on explicit quality dimensions, then pressure-test the strongest few before committing
Issue-Position-Argument mapping Turn a tangled multi-question deliberation or transcript into a typed map - the open issues, the rival positions on each, and the arguments for and against

Browse them on the live site or in _workflows/.

🧭 Find your way in

If you want to... Start here
Get unstuck or decide, now The Framework Advisor - describe your situation, get a plan
See real decisions worked prompt-to-artifact (incl. the cross-library tfs -> pm-skills handoff) The Showcase
See one quick worked example of every framework The Samples shelf
Browse by the job you need done Explore by job
See only the strong-evidence methods Explore by evidence
Filter by your situation, live The interactive chooser
Learn good thinking, beginner to advanced The learning tracks
Check the claims The evidence and bibliography
Build with or extend the library docs/ and docs/contributing.md

📖 Documentation

🥇 Conformance: what advanced (Gold) tier means

This plugin is built to the agent-skills-toolkit Advanced Skill Library Standard, which grades a skill library on three tiers. Each tier includes everything below it:

Tier Name What it certifies
🥇 Advanced (Gold) The plugin proves itself - it ships CI that runs the Standard's own validators against it and passes (self-hosting), generates its INDEX.md and native manifests from a single authored source, and maintains release notes and a deprecation policy.
🥈 Convergent (Silver) The plugin declares its agent targets and emits each higher-order component (commands, workflows, chain contracts) correctly for both Claude Code and Codex, with a manifest that matches what is on disk.
🥉 Universal (Bronze) The skills are portable - valid frontmatter, an `AGENTS