@oriro/orirocli 0.1.0

package/skills/science/sci-method/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 ORIRO.ai
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/skills/science/sci-method/SKILL.md

@@ -0,0 +1,77 @@
+---
+watermark: ORIRO
+disable-model-invocation: true
+name: sci-method
+provider: ORIRO.ai
+copyright: Copyright (c) 2026 ORIRO.ai
+description: >
+  Scientific method — hypothesis formation, experimental design, controls, variables, reproducibility, and the philosophy of science.
+
+
+
+
+
+
+  Sources: OpenStax textbooks (CC-BY), Khan Academy, MIT OCW, NOAA, IPCC reports.
+---
+
+# Scientific Method
+
+## The process
+
+1. **Observation:** Notice a phenomenon. Ask a question.
+2. **Background research:** What is already known?
+3. **Hypothesis:** Testable prediction that could be wrong.
+4. **Experiment design:** Controls, variables, measurement plan.
+5. **Data collection:** Careful, systematic, recorded.
+6. **Analysis:** Statistical analysis of data.
+7. **Conclusion:** Does data support the hypothesis? What are the limitations?
+8. **Communication:** Share findings for scrutiny and replication.
+
+## Hypothesis criteria
+
+Good hypotheses are **falsifiable** — they make predictions that could be wrong.
+"God exists" — not falsifiable, not scientific.
+"Caffeine increases short-term memory performance" — falsifiable, testable.
+
+Null hypothesis (H₀): No effect. "Caffeine has no effect on memory."
+Alternative hypothesis (H₁): Effect exists. "Caffeine improves memory."
+Statistical tests seek to reject the null hypothesis.
+
+## Experimental design
+
+**Independent variable:** What you manipulate. "Amount of caffeine consumed."
+**Dependent variable:** What you measure. "Score on memory test."
+**Control variable:** What you hold constant. "Age of participants, time of day, type of test."
+**Control group:** Receives no treatment. Used as baseline comparison.
+**Experimental group:** Receives the treatment.
+
+**Randomization:** Randomly assign participants to groups to eliminate selection bias.
+**Blinding:** Single-blind = participants don't know which group. Double-blind = participants AND experimenter don't know.
+**Sample size:** Larger = more statistical power = more reliable results.
+**Replication:** Repeat experiment. Results that can't be replicated are suspicious.
+
+## Types of studies
+
+**Observational:** Observe without manipulation. Can establish association, NOT causation.
+**Experimental:** Manipulate variable. Can establish causation IF properly controlled.
+**Longitudinal:** Follow same subjects over time. Costly but powerful for development/aging studies.
+**Cross-sectional:** Snapshot of many subjects at one time. Faster but confounded by cohort effects.
+
+## Scientific consensus vs. controversy
+
+**Scientific consensus:** Overwhelming agreement among experts based on evidence.
+Climate change, vaccine safety, evolution — scientific consensus exists on all.
+**Legitimate controversy:** Active scientific debate with evidence on multiple sides.
+**Manufactured controversy:** Controversy created by non-scientific actors to confuse public (tobacco industry playbook, fossil fuel industry on climate change).
+Distinguish: Is the disagreement among scientists or between scientists and outside interests?
+
+## Evaluating scientific claims
+
+**Peer review:** Independent experts evaluate before publication. Not perfect but essential filter.
+**Reproducibility crisis:** Many published findings don't replicate. Pre-registration, open data, larger samples help.
+**Effect size matters.** Statistically significant ≠ practically significant.
+**Funding source:** Can create bias. Not disqualifying but note it.
+**Extraordinary claims require extraordinary evidence** (Sagan).
+
+Sources: MIT OCW philosophy of science, OpenStax Biology (CC-BY), Nature "reproducibility crisis" coverage

package/skills/science/sci-neuroscience/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 ORIRO.ai
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/skills/science/sci-neuroscience/SKILL.md

@@ -0,0 +1,79 @@
+---
+watermark: ORIRO
+disable-model-invocation: true
+name: sci-neuroscience
+provider: ORIRO.ai
+copyright: Copyright (c) 2026 ORIRO.ai
+description: >
+  Neuroscience fundamentals — brain structure, neural function, memory, learning,
+  neuroplasticity, consciousness, and applied neuroscience. Activate for questions
+  about how the brain works, memory, learning, neurological conditions, or any
+  neuroscience question. Sources: NIH Brain Research, OpenStax Anatomy and
+  Physiology, Society for Neuroscience.
+---
+
+# Neuroscience Fundamentals
+
+## Brain structure
+
+**Brainstem:** Medulla, pons, midbrain. Basic life functions (breathing, heart rate, sleep). Oldest evolutionary structure.
+**Cerebellum:** "Little brain." Motor coordination, balance, procedural learning. Doesn't initiate movement — refines it.
+**Limbic system:** Emotion and memory. Key structures:
+
+- Amygdala: Fear, emotional memory, threat detection
+- Hippocampus: Memory formation, spatial navigation
+- Hypothalamus: Homeostasis, hunger, thirst, circadian rhythm, hormones
+  **Cerebral cortex:** Outer gray matter. Higher cognition. Four lobes:
+- Frontal: Planning, decision-making, personality, voluntary movement
+- Parietal: Sensory processing, spatial awareness
+- Temporal: Auditory processing, language (Wernicke's area), face recognition
+- Occipital: Visual processing
+
+## Neurons and synapses
+
+**Neuron structure:** Cell body → Dendrites (receive signals) → Axon (transmit signals) → Axon terminals (release neurotransmitters).
+**Myelin sheath:** Insulation that speeds signal transmission. MS (multiple sclerosis) = demyelination.
+**Action potential:** All-or-nothing electrochemical signal. Threshold: ~-55mV. Fires or doesn't fire.
+**Synapse:** Gap between neurons. Electrical signal → neurotransmitter release → receptor binding → new electrical signal.
+
+**Major neurotransmitters:**
+Dopamine: Reward, motivation, movement. Deficiency: Parkinson's. Excess: Psychosis. Addictive drugs hijack this system.
+Serotonin: Mood, sleep, appetite, digestion. Low levels associated with depression. SSRIs increase availability.
+GABA: Inhibitory. "The brake pedal." Alcohol, benzos, and sleeping pills enhance GABA signaling.
+Glutamate: Excitatory. Most common in brain. Learning and memory. Excess = excitotoxicity.
+Norepinephrine: Arousal, fight-or-flight. Focus and attention.
+Acetylcholine: Memory, learning, muscle control. Deficiency: Alzheimer's disease.
+
+## Memory systems
+
+**Working memory:** Active, short-term. ~7 items, ~15-30 seconds without rehearsal. Prefrontal cortex.
+**Long-term memory:**
+Explicit/declarative:
+
+- Episodic: Personal events ("my birthday last year"). Hippocampus critical.
+- Semantic: Facts ("Paris is the capital of France"). Temporal/frontal.
+  Implicit/non-declarative:
+- Procedural: Skills (riding a bike). Basal ganglia, cerebellum.
+- Priming: Unconscious memory influences. Various regions.
+- Conditioning: Associative learning. Amygdala (fear), cerebellum.
+
+**Memory formation:**
+Encoding → Consolidation (sleep essential) → Retrieval.
+LTP (Long-term potentiation): "Neurons that fire together wire together" (Hebb). Synapses strengthen with repeated co-activation.
+
+## Neuroplasticity
+
+The brain changes structure and function in response to experience throughout life.
+**Critical periods:** Early childhood — heightened plasticity. Language, vision, social development.
+**Adult plasticity:** Less dramatic but real. Hippocampal neurogenesis (new neurons) in adults demonstrated.
+**Use it or lose it:** Synaptic pruning — unused connections eliminated. Explains importance of learning and mental engagement.
+**Trauma and plasticity:** PTSD involves altered amygdala-prefrontal balance. Therapy can normalize.
+
+## Applied neuroscience for performance
+
+**Sleep:** Memory consolidation requires sleep. SWS replays episodic memories. REM processes emotional memories. Learning without sleep → far less retained.
+**Exercise:** Increases BDNF (brain-derived neurotrophic factor) — "Miracle-Gro for the brain." Improves memory, learning, mood, neurogenesis.
+**Stress:** Acute stress: enhances attention and memory. Chronic stress: damages hippocampus, impairs prefrontal function.
+**Mindfulness:** Increases gray matter density in prefrontal cortex and insula. Reduces amygdala reactivity. Not woo — documented neurological changes.
+
+Sources: NIH Brain Research (brain.nih.gov — free), Society for Neuroscience (sfn.org — free resources), OpenStax Anatomy and Physiology (CC-BY), Eagleman "The Brain" (principles)

package/skills/science/sci-physics/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 ORIRO.ai
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/skills/science/sci-physics/SKILL.md

@@ -0,0 +1,78 @@
+---
+watermark: ORIRO
+disable-model-invocation: true
+name: sci-physics
+provider: ORIRO.ai
+copyright: Copyright (c) 2026 ORIRO.ai
+description: >
+
+
+
+
+  Physics fundamentals — mechanics, thermodynamics, electricity, waves, optics, and modern physics concepts.
+
+
+  Sources: OpenStax textbooks (CC-BY), Khan Academy, MIT OCW, NOAA, IPCC reports.
+---
+
+# Physics Fundamentals
+
+## Classical mechanics
+
+**Newton's Laws:**
+1st: Object at rest stays at rest; object in motion stays in motion — unless acted on by external force.
+2nd: F = ma (Force = mass × acceleration). Net force, not individual forces.
+3rd: For every action, equal and opposite reaction.
+
+**Kinematics (constant acceleration):**
+v = v₀ + at
+x = x₀ + v₀t + ½at²
+v² = v₀² + 2a(x - x₀)
+Free fall: a = g = 9.81 m/s² downward.
+
+**Energy:**
+Kinetic energy: KE = ½mv²
+Potential energy (gravitational): PE = mgh
+Conservation of energy: Total energy constant in closed system.
+Work: W = F·d·cos(θ). Work-energy theorem: W_net = ΔKE.
+
+**Momentum:** p = mv. Conservation: p_initial = p_final in closed system.
+Elastic collision: KE conserved. Inelastic: KE not conserved.
+
+## Thermodynamics
+
+**Zeroth Law:** If A is in thermal equilibrium with B, and B with C, then A is in equilibrium with C. (Defines temperature.)
+**First Law:** ΔU = Q - W. Internal energy change = heat added - work done by system.
+**Second Law:** Entropy of isolated system never decreases. Heat flows spontaneously from hot to cold.
+**Third Law:** Entropy approaches zero as temperature approaches absolute zero.
+**Absolute zero:** 0 K = -273.15°C. Cannot be reached in practice.
+
+## Electricity and magnetism
+
+**Coulomb's Law:** F = k(q₁q₂)/r². Attraction or repulsion depending on sign.
+**Electric field:** E = F/q. Direction of force on positive test charge.
+**Ohm's Law:** V = IR. (See electrical engineering skill for circuits.)
+**Faraday's Law:** Changing magnetic flux induces EMF. Foundation of electric generators and transformers.
+**Maxwell's equations:** Four equations that describe all classical electromagnetism. Also predicted light as electromagnetic waves.
+
+**Electromagnetic spectrum:**
+Radio < Microwave < Infrared < **Visible light** < Ultraviolet < X-ray < Gamma ray.
+All travel at c = 3×10⁸ m/s in vacuum.
+E = hf = hc/λ (energy of photon).
+
+## Waves and optics
+
+**Wave properties:** Wavelength (λ), frequency (f), amplitude, speed. v = fλ.
+**Interference:** Constructive (same phase, amplitude adds). Destructive (opposite phase, amplitude cancels).
+**Doppler effect:** Observed frequency shifts with relative motion. Source approaching → higher pitch. Source receding → lower pitch. Applies to light (redshift/blueshift).
+**Snell's Law:** n₁sin(θ₁) = n₂sin(θ₂). Light bends at interface between media.
+Total internal reflection: If angle of incidence > critical angle, light reflects instead of refracts. Foundation of fiber optics.
+
+## Modern physics
+
+**Special relativity:** Speed of light constant for all observers. At relativistic speeds: time dilation, length contraction, mass increase.
+E = mc²: Mass-energy equivalence.
+**Quantum mechanics:** Energy is quantized. Particles have wave-like properties. Heisenberg uncertainty principle: ΔxΔp ≥ ℏ/2 — cannot simultaneously know position and momentum precisely.
+**Nuclear physics:** Strong nuclear force binds protons and neutrons. Fission: heavy nuclei split, releases energy. Fusion: light nuclei combine, releases even more energy per mass.
+
+Sources: OpenStax University Physics (CC-BY), MIT OCW 8.01/8.02, Khan Academy Physics

package/skills/science/sci-research-methods/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 ORIRO.ai
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/skills/science/sci-research-methods/SKILL.md

@@ -0,0 +1,83 @@
+---
+watermark: ORIRO
+disable-model-invocation: true
+name: sci-research-methods
+provider: ORIRO.ai
+copyright: Copyright (c) 2026 ORIRO.ai
+description: >
+
+
+
+
+
+
+  Research methods — study design, sampling, data collection, literature review, academic writing, and research ethics.
+  Sources: OpenStax textbooks (CC-BY), Khan Academy, MIT OCW, NOAA, IPCC reports.
+---
+
+# Research Methods
+
+## Research design hierarchy
+
+**Systematic reviews/meta-analyses:** Synthesize all studies on a question. Highest level of evidence for most questions.
+**Randomized controlled trials:** Experimental design with randomization. High internal validity.
+**Cohort studies:** Follow people over time. Can establish temporal sequence.
+**Case-control studies:** Compare exposed vs. unexposed, looking backward.
+**Cross-sectional:** Snapshot. Good for prevalence; cannot establish causation.
+**Case reports/expert opinion:** Lowest evidence. Hypothesis-generating.
+
+## Sampling
+
+**Probability sampling (scientific):** Every member of population has known probability of selection.
+Simple random: Every member equal probability. Gold standard for generalization.
+Stratified: Divide into groups, sample from each. Ensures representation.
+Cluster: Sample groups, then individuals within groups. More practical for large populations.
+Systematic: Every nth person from a list.
+
+**Non-probability sampling (practical):**
+Convenience: Whoever is available. High bias. Limits generalizability.
+Purposive: Deliberately select specific types. Qualitative research.
+Snowball: Participants recruit other participants. Social networks, hidden populations.
+
+**Sample size determination:** Based on effect size, desired power (0.80), and significance level (0.05). Use G\*Power or similar calculators.
+
+## Literature review
+
+**Systematic vs. narrative:**
+Systematic: Pre-specified search strategy, inclusion/exclusion criteria, documented. Reproducible.
+Narrative: Selectively reviews literature based on author's knowledge. Prone to bias.
+
+**Search strategy:**
+Databases: PubMed (biomedicine), PsycINFO (psychology), Scopus, Web of Science, JSTOR, Google Scholar.
+Boolean operators: AND (narrow), OR (expand), NOT (exclude).
+MeSH terms: Medical Subject Headings — controlled vocabulary for PubMed.
+
+**Critical appraisal:** Don't accept published studies uncritically.
+Check: Is the study design appropriate for the question? Was randomization done correctly? Was blinding appropriate? Was the sample size adequate? Were confounders addressed? Do conclusions match the data?
+
+**PICO search strategy:** Population + Intervention + Comparison + Outcome → generates focused search terms.
+
+## Academic writing
+
+**IMRaD structure:** Introduction, Methods, Results, and Discussion.
+Introduction: What is known? What is the gap? What is the research question?
+Methods: Detailed enough to replicate. Study design, participants, measurements, analysis plan.
+Results: What was found. Data, statistics, tables, figures. No interpretation here.
+Discussion: What does it mean? Compare to prior literature. Limitations. Implications.
+
+**Writing principles:**
+Precision: Use exact terms. Avoid weasel words (may, perhaps, seems).
+Economy: One idea per sentence. Delete unnecessary words.
+Active vs. passive: "We conducted..." vs "The study was conducted..." Active is clearer.
+Citation management: Zotero (free), Mendeley, EndnoteRelevant.
+
+## Research ethics
+
+**Informed consent:** Participants must understand: what the study involves, risks and benefits, right to withdraw, how data will be used.
+**Confidentiality vs. anonymity:** Confidentiality = identifiable but protected; Anonymity = cannot be identified.
+**IRB/Ethics review:** Required for research involving human subjects.
+**Research misconduct:** Fabrication (making up data), Falsification (manipulating data), Plagiarism (taking credit for others' work). Career-ending and potentially criminal.
+**Publication bias:** Positive results more likely to be published. Contributes to replication crisis. Pre-registration helps.
+
+Sources: OpenStax Research Methods (CC-BY), Cochrane handbook (free, cochrane.org),
+Creswell Research Design (principles), APA Publication Manual (principles)

package/skills/science/sci-statistics/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 ORIRO.ai
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/skills/science/sci-statistics/SKILL.md

@@ -0,0 +1,249 @@
+---
+watermark: ORIRO
+name: sci-statistics
+provider: ORIRO.ai
+copyright: Copyright (c) 2026 ORIRO.ai
+description: >
+  Statistics and data analysis — probability, distributions, hypothesis testing,
+  regression, study design, and interpreting results. Activate for questions
+  about data analysis, p-values, confidence intervals, statistical significance,
+  sample size, surveys, A/B testing, correlations, or any statistics question.
+  Sources: OpenStax Statistics (CC-BY), Khan Academy Statistics, MIT OCW
+  Statistics for Applications, ASA statistical guidelines.
+---
+
+# Statistics
+
+## Descriptive statistics
+
+### Measures of center
+
+**Mean:** Sum of all values ÷ number of values. Sensitive to outliers.
+**Median:** Middle value when sorted. Robust to outliers. Better for skewed data.
+**Mode:** Most frequent value. Useful for categorical data.
+
+When to use which:
+
+- Symmetric distribution: mean ≈ median — use either.
+- Skewed distribution: median more representative than mean.
+- Example: Income distribution is right-skewed. Median income is more informative than mean.
+
+### Measures of spread
+
+**Range:** Max - Min. Sensitive to outliers.
+**Variance:** Average squared deviation from the mean.
+**Standard deviation (SD):** Square root of variance. Same units as the data.
+**IQR (Interquartile Range):** Q3 - Q1. Middle 50% of data. Robust to outliers.
+
+### The 68-95-99.7 rule (normal distribution)
+
+In a normally distributed dataset:
+
+- 68% of data falls within ±1 SD of the mean
+- 95% within ±2 SD
+- 99.7% within ±3 SD
+
+**Outlier identification:** Values more than 1.5×IQR beyond Q1 or Q3 (box plot method). Or > 3 SD from mean.
+
+---
+
+## Probability fundamentals
+
+### Basic rules
+
+P(A or B) = P(A) + P(B) - P(A and B)
+P(A and B) = P(A) × P(B|A) — conditional probability
+If A and B are independent: P(A and B) = P(A) × P(B)
+
+### Bayes' theorem
+
+P(A|B) = [P(B|A) × P(A)] / P(B)
+
+**Medical testing example:**
+
+- Disease prevalence: 1% (P(disease) = 0.01)
+- Test sensitivity: 99% (P(positive|disease) = 0.99)
+- Test specificity: 95% (P(negative|no disease) = 0.95)
+- P(positive|no disease) = 0.05
+
+P(disease|positive test) = (0.99 × 0.01) / [(0.99 × 0.01) + (0.05 × 0.99)]
+= 0.0099 / (0.0099 + 0.0495) = 0.0099 / 0.0594 ≈ 16.7%
+
+Even with a positive test, only 16.7% chance of having the disease.
+This is why mass screening of low-prevalence conditions generates mostly false positives.
+
+---
+
+## Statistical inference
+
+### Sampling
+
+**Population:** The entire group you want to know about.
+**Sample:** The subset you actually measure.
+**Random sampling:** Every member of population has equal probability of selection.
+**Sample size:** Larger = more precise estimates; diminishing returns.
+
+**Sampling errors (types of bias):**
+
+- Selection bias: sample not representative of population
+- Non-response bias: non-respondents differ from respondents
+- Volunteer bias: volunteers differ from non-volunteers
+
+### Confidence intervals
+
+A 95% confidence interval means: if we repeated the study 100 times, 95 of those
+intervals would contain the true population parameter.
+
+**It does NOT mean:** There's a 95% probability the true value is in this specific interval.
+
+**Interpreting width:**
+Wide CI = high uncertainty (small sample, high variability)
+Narrow CI = more precision (large sample, low variability)
+
+### Hypothesis testing
+
+**Null hypothesis (H₀):** The default assumption (usually "no effect" or "no difference").
+**Alternative hypothesis (H₁):** What you're trying to show evidence for.
+
+**P-value:** Probability of observing results at least as extreme as yours, IF the null hypothesis were true.
+
+**Common misinterpretation:** p-value is NOT the probability that the null hypothesis is true.
+
+**Significance level (α):** Threshold for rejecting null hypothesis.
+Commonly α = 0.05 (5% false positive rate).
+High-stakes fields use α = 0.01 or 0.001.
+
+**Statistical power:** Probability of detecting a real effect if one exists.
+Convention: power ≥ 0.80 (80% chance of detecting true effect).
+Low power → high false negative rate.
+
+**Type I error (α):** Reject null when it's actually true. (False positive)
+**Type II error (β):** Fail to reject null when it's actually false. (False negative)
+
+### What p < 0.05 actually means
+
+Statistical significance ≠ practical significance.
+With a large enough sample, even trivially small effects are "statistically significant."
+
+**Always ask:** Effect size — how large is the difference in practical terms?
+**Cohen's d:** Effect size for means comparison.
+d = 0.2 (small), 0.5 (medium), 0.8 (large)
+
+---
+
+## Common statistical tests
+
+### Choosing the right test
+
+| Data type              | Two groups             | Multiple groups | Correlation |
+| ---------------------- | ---------------------- | --------------- | ----------- |
+| Continuous, normal     | t-test                 | ANOVA           | Pearson r   |
+| Continuous, non-normal | Mann-Whitney U         | Kruskal-Wallis  | Spearman ρ  |
+| Categorical            | Chi-square             | Chi-square      | Chi-square  |
+| Proportions            | Z-test for proportions | —               | —           |
+
+### t-test (compare two group means)
+
+One-sample: Does sample mean differ from a known value?
+Independent samples: Do two groups differ from each other?
+Paired: Does the same group change before vs. after?
+
+### Chi-square test (categorical data)
+
+Tests if observed frequencies differ from expected.
+Or tests independence between two categorical variables.
+
+### ANOVA (Analysis of Variance)
+
+Compares means of three or more groups simultaneously.
+If significant: post-hoc tests (Tukey, Bonferroni) identify which groups differ.
+
+### Correlation
+
+**Pearson r:** Linear relationship between two continuous variables.
+r = +1: perfect positive correlation
+r = 0: no linear relationship
+r = -1: perfect negative correlation
+
+**Correlation ≠ causation.** Classic error in statistics communication.
+Both could be caused by a third variable (confounding).
+Spurious correlations: ice cream sales and drowning deaths (both caused by summer).
+
+---
+
+## Regression
+
+### Simple linear regression
+
+y = β₀ + β₁x + ε
+y = outcome (dependent variable)
+x = predictor (independent variable)
+β₁ = slope (change in y per unit change in x)
+β₀ = intercept
+
+**R²:** Proportion of variance in y explained by x.
+R² = 0.70 means: 70% of variation in the outcome is explained by the model.
+
+### Multiple regression
+
+y = β₀ + β₁x₁ + β₂x₂ + ... + βₙxₙ + ε
+Each β interpreted as effect of that variable, holding all others constant.
+Critical for controlling confounders.
+
+### Assumptions of linear regression
+
+1. Linearity: relationship between x and y is linear
+2. Independence: observations are independent
+3. Homoscedasticity: variance of residuals is constant
+4. Normality of residuals: residuals approximately normally distributed
+5. No perfect multicollinearity
+
+---
+
+## Study design
+
+### Observational studies
+
+**Cross-sectional:** Snapshot at one point in time. Good for prevalence.
+**Case-control:** Compare people with/without outcome; look backward at exposures.
+**Cohort:** Follow people over time; compare those with/without exposure.
+Prospective: follow forward. Retrospective: look at historical records.
+
+### Experimental studies
+
+**Randomized controlled trial (RCT):** Gold standard.
+Random assignment to treatment/control eliminates selection bias.
+
+**Blinding:**
+Single-blind: participants don't know which group.
+Double-blind: participants AND investigators don't know.
+Reduces performance bias, detection bias, and placebo effects.
+
+### A/B Testing (web analytics context)
+
+Randomize users to version A or version B.
+Collect enough data to reach statistical significance.
+
+**Common errors:**
+
+- Peeking: stopping test early when you see significance (inflates false positives)
+- Multiple testing: testing many variations simultaneously inflates false positive rate
+- Novelty effect: new design performs well initially but regresses
+
+**Minimum sample size:**
+n ≥ (Z*α/2 + Z*β)² × 2σ² / Δ²
+For binary outcomes: use online calculators with power = 0.80, α = 0.05.
+
+---
+
+## Simpson's Paradox
+
+A trend appears in subgroups but disappears or reverses when combined.
+Example: Treatment A has better survival rates in both severe and mild cases,
+but Treatment B has a better overall survival rate because more mild cases got Treatment B.
+
+**Solution:** Always stratify by relevant confounders. Never trust aggregated data alone.
+
+Sources: OpenStax Statistics (CC-BY, openstax.org), Khan Academy Statistics (CC-BY-NC-SA),
+MIT OCW 18.650 Statistics for Applications, American Statistical Association
+guidelines on p-values (2016 statement)