Cognitive Mode-Switch Words

Words that don't describe what you want — they rewire how you approach the problem.

Quick Reference: Word → Behavior

When the user uses these words, activate the corresponding thinking mode:

Depth & Rigor

Word	Thinking Mode	Invoke Skill
Rigorous	Proof-like thinking — every step must earn its place	`rigorous-thinking`
First-principles	Derive from ground truth, strip borrowed reasoning	`rigorous-thinking`
Mechanistic	Demand causal chains: "A causes B because..."	`rigorous-thinking`
Axiomatic	Each claim follows logically from stated premises	`rigorous-thinking`
Falsifiable	Flip to "what would prove this wrong?"	`rigorous-thinking`
Steelman	Construct strongest possible version of opposing view	`rigorous-thinking`, `rigorous-exchange`
Non-trivial	Skip basics, assume expertise, go deeper	—
Exhaustive	Complete coverage, no blind spots, breadth before depth	—
Necessary and sufficient	Exact boundaries — no more, no less	`claim-validator`
Load-bearing	Identify the single assumption everything depends on	`rigorous-thinking`
Dispositive	Find the one fact that settles the debate	`rigorous-thinking`
Ultra think hard	Maximum cognitive effort, no shortcuts anywhere	`rigorous-thinking`

Creative Reframing

Word	Thinking Mode	Invoke Skill
Reimagine	Break fixed mental models — clean-slate vision	`ideacritic`
Orthogonal	Think perpendicular to the current frame	`ideacritic`
Contrarian	Deliberate inversion of the popular view	`ideacritic`
Subversive	Question the premise itself, not just disagree	`ideacritic`
Invert	Instead of "how to succeed," ask "how to guarantee failure"	`inversion-analysis`
Counterfactual	"What would have happened if X didn't occur?"	`inversion-analysis`
Zero-based	Start from scratch — if you had none of this, what would you build?	`ideacritic`
Provocative	Bold, non-consensus ideas you'd never say in a meeting	`ideacritic`
Transmute	Convert a weakness into a strength	`ideacritic`
Reframe	Re-examine whether you're solving the right problem	`ideacritic`
Degenerate case	Push variables to 0, 1, or infinity to reveal hidden structure	—
Adjacent possible	Constrain creativity to what's buildable from current position	—

Precision & Mathematical

Word	Thinking Mode	Invoke Skill
Probabilities/Distribution	Full statistical reasoning — likelihoods, not certainties	`claim-validator`
Bounded	Force upper and lower limits on any assertion	`claim-validator`, `unit-economics-validator`
Monotonic	Does this always increase/decrease, or are there reversals?	—
Asymptotic	What happens as scale approaches infinity?	—
Invariant	Identify the constant in a sea of variables	—
Order of magnitude	Calibration without false precision	`unit-economics-validator`
Sensitivity analysis	Which variable, if wrong by 2x, breaks the conclusion?	`unit-economics-validator`
Convex	Is the payoff shape concave (diminishing) or convex (accelerating)?	—
Marginal	What does one more unit buy you? Not average, incremental.	`pricing-strategy`
Dimensionality	How many independent axes actually matter?	—
Closed-form	Demand a computable formula, not hand-waving	—

Tradeoff & Decision

Word	Thinking Mode	Invoke Skill
Sacrifice	Name what you're willing to lose — make tradeoffs explicit	`operator-playbook`
Irreversible	Can we undo this or not? One-way vs two-way doors	`operator-playbook`
Binding constraint	Identify the single bottleneck	`systems-decompose`
Pareto	What 20% of effort gets 80% of the result?	`operator-playbook`
Second-order effects	What happens after the thing happens? Ripple effects.	`systems-review`
Opportunity cost	Is this better than the next best alternative?	`operator-playbook`
Regret minimization	Which choice do you regret least looking back from 10 years out?	`operator-playbook`
Non-negotiable	What will you absolutely not bend on?	—
Asymmetric	Where upside and downside are mismatched	—
Reversibility gradient	Not binary — place decision on the spectrum	—

Quality & Craft

Word	Thinking Mode	Invoke Skill
Elegant	Simplicity, completeness, and surprise in one solution	`code-review`, `software-architecture`
Idiomatic	Domain-native solutions, not generic ones	`code-review`
Canonical	The accepted best practice, not improvisation	`code-review`
Production-grade	Error handling, edge cases, monitoring, logging	`code-review`, `security-review`
Battle-tested	Proven approaches over clever theory	`software-architecture`
Minimalist	Remove everything that isn't essential	`simplify`
Composable	Modular design where pieces can be mixed and matched	`software-architecture`
Tasteful	Aesthetic reasoning — what feels right to an expert?	—
Crisp	Precision in language — every word must earn its place	`deslop`

Systems Thinking

Word	Thinking Mode	Invoke Skill
Emergent	What arises from interactions that no individual part possesses?	`systems-decompose`
Cascading	Maps domino effects through a system	`systems-review`
Equilibrium	Is this equilibrium stable or unstable? What disrupts it?	—
Phase transition	At what threshold does the system fundamentally shift?	—
Feedback loop	Is this positive (amplifying) or negative (stabilizing)?	`systems-decompose`
Attractor	Where does this naturally converge?	—
Entropy	What's degrading over time? What maintenance prevents collapse?	`techdebt`

Epistemic Calibration

Word	Thinking Mode	Invoke Skill
Bayesian	Explicit priors, evidence weighting, posterior updates	`rigorous-thinking`
Calibrated	Honest probability assignment to claims	`rigorous-thinking`
Crux	Single disagreement that, if settled, resolves everything	`rigorous-exchange`
Pre-mortem	Assume this failed. Now explain why.	`inversion-analysis`
Evidence-weighted	Which signals carry the most weight and why?	`rigorous-thinking`
Adversarial	Assume someone is actively trying to break this	`security-review`
Sparse	Strip away everything except what carries information	`rigorous-thinking`
Steel-thread	Thinnest possible end-to-end proof of concept	—
Epistemic humility	Honest "here's what I don't know" disclosure	`rigorous-thinking`
Signal-to-noise	What fraction of this information changes decisions?	`rigorous-thinking`

Power Combos

The highest-leverage technique: pair a depth word with a constraint word.

Combo	Effect	Invoke
Rigorous + Elegant	Airtight reasoning that's also beautiful and simple	`rigorous-thinking` + `code-review`
Reimagine + First-principles	Clean-slate vision built from ground truth	`ideacritic` + `rigorous-thinking`
Ultra think hard + Sacrifice	Maximum depth applied to what must be cut	`rigorous-thinking` + `operator-playbook`
Falsifiable + Bounded	Testable claims with defined ranges	`rigorous-thinking` + `claim-validator`
Adversarial + Exhaustive	Complete threat surface mapping	`security-review`
Contrarian + Evidence-weighted	Bold takes backed by real data	`ideacritic` + `rigorous-thinking`
Invert + Pre-mortem	Failure-first scenario planning	`inversion-analysis`
Sparse + Dispositive	Cut to the one thing that decides everything	`rigorous-thinking`
Distribution + Sensitivity analysis	Full outcome range with fragility mapping	`unit-economics-validator`

Usage

This skill auto-invokes when the user uses trigger words. It then:

Activates the thinking mode corresponding to the word(s)
Invokes relevant skills if deeper analysis is needed
Applies the cognitive constraint to all reasoning in the response

Examples

User: "Be rigorous: why won't this architecture scale?" → Invoke rigorous-thinking, apply proof-like reasoning

User: "Give me the contrarian take on our pricing." → Invoke ideacritic, deliberately invert the popular view

User: "Ultra think hard + sacrifice: what do we cut?" → Maximum cognitive effort applied to naming tradeoffs

User: "Sensitivity analysis on our unit economics." → Invoke unit-economics-validator, identify fragile assumptions

The Meta-Pattern

Words that constrain the solution space in a non-obvious way always outperform words that merely describe desired quality.

❌ "Good" is useless
✅ "Elegant" rewires the entire approach
✅ "Falsifiable" flips from prove-right to prove-wrong
✅ "Sparse" strips away noise before answering

When you hear a trigger word, change HOW you think, not just WHAT you output.

ai-lingo