feat: v0.4.0-beta — self-hosted registry + new skills

.claude/skills/first-principles-thinking/references/bias-and-debiasing.md

@@ -0,0 +1,124 @@
+# Cognitive Biases That Corrupt First Principles Thinking
+
+First-principles reasoning demands that you strip away assumptions and reason from ground truths. But your brain is wired to do the opposite — it takes shortcuts, clings to patterns, and avoids the discomfort of uncertainty. These 12 biases are the most dangerous enemies of first-principles thinking because they disguise assumptions as truths.
+
+Source: think-better project.
+
+---
+
+## The 12 Biases
+
+### 1. Confirmation Bias
+
+**What it does**: You seek, interpret, and remember evidence that supports your existing beliefs.
+**How it corrupts FP thinking**: When decomposing to ground truths, you unconsciously "discover" foundations that happen to support your preferred conclusion. The analysis feels rigorous but the destination was predetermined. This is the single most dangerous bias for FP reasoning because it makes bad analysis feel airtight.
+**Debiasing strategy**: Actively seek disconfirming evidence first. Before building up from your ground truths, spend 10 minutes trying to prove them wrong. Write down what evidence would change your mind — then go look for it.
+**Key question**: "Am I looking for truth, or am I looking for permission?"
+
+### 2. Anchoring
+
+**What it does**: You over-rely on the first piece of information you encounter, and all subsequent reasoning adjusts from that anchor.
+**How it corrupts FP thinking**: The conventional approach becomes your mental anchor even when you are trying to think from scratch. Your "first principles" solution ends up suspiciously close to the status quo with minor adjustments. The anchor is invisible — you genuinely believe you started from zero.
+**Debiasing strategy**: Generate 3 alternatives before evaluating any single one. Force yourself to produce options that are structurally different, not incremental variations. Deliberately include one option that would make your industry peers uncomfortable.
+**Key question**: "Is my 'from scratch' solution just the existing approach with cosmetic changes?"
+
+### 3. Sunk Cost Fallacy
+
+**What it does**: You continue investing in something because of past investment, not future value.
+**How it corrupts FP thinking**: Legacy decisions feel like ground truths because so much was invested in them. "We use X because we spent two years building it" masquerades as "X is the right foundation." The more you have invested, the harder it is to see the investment as irrelevant to the forward-looking decision.
+**Debiasing strategy**: Ask "If we were starting from zero today with everything we now know, would we make this same choice?" If no, the sunk cost is distorting your reasoning. Treat all past investments as write-offs when evaluating the path forward.
+**Key question**: "Would I choose this if I hadn't already invested in it?"
+
+### 4. Status Quo Bias
+
+**What it does**: You prefer the current state of affairs simply because it is current.
+**How it corrupts FP thinking**: "How it works now" feels like a constraint when it is actually a choice. You unconsciously preserve existing structures and call them fundamentals. This is particularly insidious in organizations where "the way we do things" has been codified into process documents that look like laws.
+**Debiasing strategy**: Explicitly list everything you are treating as a constraint. For each one, ask: "Is this a law of physics, a regulation, or just how we currently do it?" Reclassify choices as choices. Be ruthless — most "constraints" are preferences.
+**Key question**: "Am I treating a choice as a constraint?"
+
+### 5. Overconfidence
+
+**What it does**: You overestimate the accuracy and completeness of your own knowledge and beliefs.
+**How it corrupts FP thinking**: You treat assumptions as verified ground truths without testing them. Your confidence in your decomposition prevents you from noticing gaps. The cleaner and more elegant your first-principles analysis feels, the more suspicious you should be — reality is usually messier than your model.
+**Debiasing strategy**: Assign a confidence percentage to each assumption in your reasoning chain. Anything below 80% needs external validation before you build on it. If every assumption is above 90%, you are almost certainly overconfident.
+**Key question**: "How would I know if this assumption were wrong?"
+
+### 6. Framing Effect
+
+**What it does**: Your decisions change depending on how information is presented, even when the underlying facts are identical.
+**How it corrupts FP thinking**: You accept the problem statement as given instead of questioning whether it is the right problem. The frame handed to you becomes an invisible constraint. In organizations, whoever writes the brief controls the conclusion — even for people doing "first principles" analysis on it.
+**Debiasing strategy**: Restate the problem in 3 structurally different ways before solving it. Change the subject, invert the goal, or shift the level of abstraction. If the problem was framed by someone else, ask what they left out of the frame.
+**Key question**: "Who framed this problem, and what did their framing exclude?"
+
+### 7. Availability Heuristic
+
+**What it does**: You judge the likelihood or importance of something by how easily examples come to mind.
+**How it corrupts FP thinking**: Recent, vivid, or emotionally charged experiences get elevated to universal truths. A single dramatic failure becomes "this never works." A recent success becomes "this always works." Your ground truths end up being a highlight reel of your recent memory, not an objective assessment.
+**Debiasing strategy**: Ask "Am I generalizing from a small or biased sample?" Seek base rates and systematic data instead of relying on memorable anecdotes. If you can name the specific example driving your belief, it is probably an anecdote, not a pattern.
+**Key question**: "Is this a pattern or just a memorable example?"
+
+### 8. Groupthink
+
+**What it does**: You conform to group consensus to maintain harmony, suppressing dissent and alternative viewpoints.
+**How it corrupts FP thinking**: Contrarian ground truths get silenced because they threaten group cohesion. The group converges on comfortable "first principles" that everyone already agrees with, defeating the entire purpose of the exercise. First-principles analysis done by a group often produces the safest possible answer, not the truest one.
+**Debiasing strategy**: Assign a devil's advocate role that rotates each session. Make it someone's explicit job to challenge every assumption, especially the ones that feel obvious. Collect written input before group discussion to prevent anchoring on the first speaker.
+**Key question**: "Is everyone agreeing because this is true, or because disagreeing is uncomfortable?"
+
+### 9. Planning Fallacy
+
+**What it does**: You systematically underestimate the time, cost, and difficulty of future actions.
+**How it corrupts FP thinking**: "Build from scratch" solutions seem easier and faster than they are. First-principles reasoning can make you overconfident about novel approaches because the analysis feels clean while the execution will not be. The gap between a correct first-principles conclusion and a successful implementation is where most FP-driven initiatives die.
+**Debiasing strategy**: Use reference class forecasting. Find 3-5 comparable projects (not your own optimistic plan) and use their actual timelines as your baseline. Multiply your time estimate by 2x for familiar work and 3x for novel work.
+**Key question**: "How long did this take the last 5 times someone tried it?"
+
+### 10. Loss Aversion
+
+**What it does**: You feel losses roughly twice as intensely as equivalent gains, making you avoid giving things up.
+**How it corrupts FP thinking**: You avoid discarding assumptions, frameworks, or prior work even when evidence says you should. Letting go of a "ground truth" feels like a loss, so you protect it past its usefulness. This bias is especially strong when the assumption was hard-won or publicly stated.
+**Debiasing strategy**: Frame decisions as "what do we gain by updating our assumptions" instead of "what do we lose by abandoning our position." Reframe every loss as a trade. Explicitly name what you gain by letting go.
+**Key question**: "Am I holding onto this because it is true, or because letting go feels like losing?"
+
+### 11. Dunning-Kruger Effect
+
+**What it does**: You overestimate your competence in areas where you have little expertise, while experts underestimate theirs.
+**How it corrupts FP thinking**: Naive first-principles analysis in an unfamiliar domain feels sufficient and elegant. You miss non-obvious constraints, edge cases, and failure modes that domain experts would catch instantly. The analysis is clean because it is incomplete. This is the classic trap for smart generalists entering a new domain.
+**Debiasing strategy**: Ask "Who has spent 10,000 hours in this domain?" Find them and pressure-test your reasoning before acting on it. Pay particular attention when your analysis contradicts expert consensus — you might be right, but the base rate says you are probably wrong.
+**Key question**: "Would a domain expert laugh at my decomposition?"
+
+### 12. Survivorship Bias
+
+**What it does**: You draw conclusions only from visible successes while ignoring the invisible failures.
+**How it corrupts FP thinking**: You build ground truths from winners' stories. "Successful companies did X, therefore X is a ground truth." But you never studied the companies that also did X and failed. Survivorship bias makes first-principles analysis converge on whatever the current winners are doing, which is the opposite of thinking from scratch.
+**Debiasing strategy**: Actively research failures in this domain. For every success story informing your reasoning, find 2 failure stories. Ask what they had in common with the successes — that commonality is probably not what caused the success.
+**Key question**: "Am I only looking at survivors?"
+
+---
+
+## Bias Categories
+
+These 12 biases cluster into 3 attack vectors against first-principles thinking:
+
+**Biases that corrupt your inputs** (what you accept as ground truths):
+- Confirmation Bias, Availability Heuristic, Survivorship Bias, Framing Effect
+
+**Biases that protect existing conclusions** (what you refuse to discard):
+- Anchoring, Sunk Cost Fallacy, Status Quo Bias, Loss Aversion
+
+**Biases that inflate your confidence** (what you think you know):
+- Overconfidence, Dunning-Kruger Effect, Planning Fallacy, Groupthink
+
+If you only have time to check one category, check the first — corrupted inputs produce confidently wrong conclusions.
+
+---
+
+## Bias Detection Checklist
+
+Run these 5 questions before finalizing any first-principles analysis:
+
+1. **Evidence check** — "Did I seek disconfirming evidence with the same energy I sought confirming evidence?"
+2. **Constraint audit** — "For each constraint in my reasoning, is it a physical law, a regulation, or just a current choice?"
+3. **Sample check** — "Are my ground truths based on systematic evidence or vivid anecdotes?"
+4. **Expertise check** — "Has someone with deep domain expertise reviewed my decomposition?"
+5. **Inversion check** — "If I argued the opposite position, what ground truths would I use — and are any of them actually stronger?"
+
+If you cannot confidently answer all 5, your first-principles analysis is likely contaminated. Go back and stress-test before building forward.