Skip to content

brycewang-stanford/Auto-Empirical-Research-Skills / skills/24-Imbad0202-academic-research-skills/deep-research

deep-research

Verified

Universal deep research agent team. 13-agent pipeline for rigorous academic research on any topic. 7 modes: full research, quick brief, paper review, lit-review, fact-check, Socratic guided research dialogue, and systematic review with optional meta-analysis. Covers research question formulation, Socratic mentoring, methodology design, systematic literature search, source verification, cross-source synthesis, risk of bias assessment, meta-analysis, APA 7.0 report compilation, editorial review, devil's advocate challenges, ethics review, and post-research literature monitoring. Triggers on: research, deep research, literature review, systematic review, meta-analysis, PRISMA, evidence synthesis, fact-check, guide my research, help me think through, 研究, 深度研究, 文獻回顧, 文獻探討, 系統性回顧, 後設分析, 事實查核, 引導我的研究, 幫我釐清, 幫我想想, 我不確定要研究什麼, 研究方向, 研究主題.

FileKindSize
SKILL.mdskill md31.8 KB
agents/bibliography_agent.mdmarkdown3.6 KB
agents/devils_advocate_agent.mdmarkdown6.1 KB
agents/editor_in_chief_agent.mdmarkdown4.9 KB
agents/ethics_review_agent.mdmarkdown6.9 KB
agents/meta_analysis_agent.mdmarkdown14.6 KB
agents/monitoring_agent.mdmarkdown9.0 KB
agents/report_compiler_agent.mdmarkdown5.4 KB
agents/research_architect_agent.mdmarkdown6.9 KB
agents/research_question_agent.mdmarkdown6.9 KB
agents/risk_of_bias_agent.mdmarkdown9.4 KB
agents/socratic_mentor_agent.mdmarkdown21.2 KB
agents/source_verification_agent.mdmarkdown6.4 KB
agents/synthesis_agent.mdmarkdown6.0 KB
examples/exploratory_research.mdmarkdown6.9 KB
examples/fact_check_mode.mdmarkdown12.9 KB
examples/handoff_to_paper.mdmarkdown11.0 KB
examples/policy_analysis.mdmarkdown9.4 KB
examples/review_mode.mdmarkdown17.3 KB
examples/socratic_guided_research.mdmarkdown22.9 KB
examples/systematic_review.mdmarkdown7.7 KB
references/apa7_style_guide.mdmarkdown4.6 KB
references/equator_reporting_guidelines.mdmarkdown14.0 KB
references/ethics_checklist.mdmarkdown10.2 KB
references/failure_paths.mdmarkdown14.8 KB
references/interdisciplinary_bridges.mdmarkdown20.9 KB
references/irb_decision_tree.mdmarkdown11.5 KB
references/literature_monitoring_strategies.mdmarkdown9.6 KB
references/logical_fallacies.mdmarkdown10.4 KB
references/methodology_patterns.mdmarkdown17.0 KB
references/mode_selection_guide.mdmarkdown16.2 KB
references/preregistration_guide.mdmarkdown11.8 KB
references/socratic_questioning_framework.mdmarkdown11.4 KB
references/source_quality_hierarchy.mdmarkdown7.2 KB
references/systematic_review_toolkit.mdmarkdown18.7 KB
templates/evidence_assessment_template.mdmarkdown3.7 KB
templates/literature_matrix_template.mdmarkdown3.7 KB
templates/preregistration_template.mdmarkdown8.5 KB
templates/prisma_protocol_template.mdmarkdown8.8 KB
templates/prisma_report_template.mdmarkdown15.8 KB
templates/research_brief_template.mdmarkdown2.1 KB

references/logical_fallacies.md

View on GitHub
# Logical Fallacies Catalog — 30+ Fallacies for Research Review

## Purpose
Reference catalog of logical fallacies commonly encountered in research. Used by the devils_advocate_agent.

## Formal Fallacies (Invalid Logical Structure)

### 1. Affirming the Consequent
**Structure**: If P then Q; Q is true; therefore P is true.
**Example**: "If a university has high research funding, it has good outcomes. This university has good outcomes. Therefore, it must have high research funding."
**Problem**: Q can have multiple causes.

### 2. Denying the Antecedent
**Structure**: If P then Q; not P; therefore not Q.
**Example**: "If enrollment increases, revenue increases. Enrollment didn't increase. Therefore, revenue didn't increase."
**Problem**: Revenue can increase from other sources.

### 3. Undistributed Middle
**Structure**: All A are B; All C are B; therefore All A are C.
**Example**: "All successful programs use technology. Our program uses technology. Therefore, our program is successful."
**Problem**: B (technology use) is shared but doesn't link A and C.

### 4. False Dilemma / False Dichotomy
**Structure**: Either A or B; not A; therefore B.
**Example**: "Either we adopt online learning completely or maintain traditional methods."
**Problem**: Many hybrid options exist.

## Informal Fallacies

### Relevance Fallacies

### 5. Ad Hominem
**Description**: Attacking the person rather than the argument.
**Research Example**: "This study's conclusions are unreliable because the author works for a for-profit university."
**Correct Approach**: Evaluate the methodology and evidence, not the author's affiliation (though COI should be noted).

### 6. Appeal to Authority
**Description**: Accepting a claim solely because an authority figure endorses it.
**Research Example**: "Published in Nature, so the findings must be valid."
**Correct Approach**: Even prestigious journals publish flawed studies. Evaluate on merit.

### 7. Appeal to Tradition
**Description**: Arguing something is correct because it has always been done that way.
**Research Example**: "This metric has been used for 30 years, so it must be the best measure."
**Correct Approach**: Evaluate whether the metric is still valid in current context.

### 8. Appeal to Novelty
**Description**: Arguing something is better because it's new.
**Research Example**: "This new framework must be superior to the established one."
**Correct Approach**: Novelty doesn't equal improvement. Compare on evidence.

### 9. Appeal to Popularity (Bandwagon)
**Description**: Arguing something is true because many people believe it.
**Research Example**: "Most researchers in the field use this method, so it must be the best."
**Correct Approach**: Popularity doesn't validate methodology. Assess independently.

### 10. Red Herring
**Description**: Introducing an irrelevant topic to divert from the argument.
**Research Example**: Responding to criticism of methodology by discussing the importance of the topic.

### Evidence Fallacies

### 11. Cherry-Picking (Selection Bias)
**Description**: Selecting evidence that supports the conclusion while ignoring contradictory evidence.
**Research Example**: Citing 5 studies that support the hypothesis while omitting 12 that don't.
**Detection**: Compare cited sources against comprehensive search results.

### 12. Confirmation Bias
**Description**: Seeking, interpreting, and remembering information that confirms pre-existing beliefs.
**Research Example**: Designing search terms that are more likely to return supportive results.
**Detection**: Check if search strategy was neutral; look for actively sought disconfirming evidence.

### 13. Survivorship Bias
**Description**: Drawing conclusions only from "survivors" (successes), ignoring those that didn't survive.
**Research Example**: "All top-ranked universities implemented X" — ignoring universities that implemented X and didn't improve.
**Detection**: Ask "what about those that failed?"

### 14. Anecdotal Evidence
**Description**: Using individual stories as proof of a general claim.
**Research Example**: "One university tripled enrollment after rebranding, so rebranding drives enrollment."
**Detection**: Is this a systematic finding or an isolated case?

### 15. Hasty Generalization
**Description**: Drawing broad conclusions from insufficient evidence.
**Research Example**: "Three case studies from Taiwan show X, therefore this applies to all Asian universities."
**Detection**: Is the sample representative? Is the generalization proportionate to the evidence?

### Causal Fallacies

### 16. Post Hoc Ergo Propter Hoc
**Description**: Assuming that because B followed A, A caused B.
**Research Example**: "After implementing the new curriculum, graduation rates improved. Therefore the curriculum caused the improvement."
**Detection**: Were there confounders? Was there a control group?

### 17. Cum Hoc Ergo Propter Hoc (Correlation ≠ Causation)
**Description**: Assuming that correlation implies causation.
**Research Example**: "Universities with more international students have higher rankings, so international students cause higher rankings."
**Detection**: Is there a plausible mechanism? Could both be caused by a third factor?

### 18. Reverse Causation
**Description**: Getting cause and effect backwards.
**Research Example**: "Good facilities attract students" when actually "student fees fund better facilities."
**Detection**: Consider temporal order and alternative causal directions.

### 19. Ecological Fallacy
**Description**: Inferring individual-level conclusions from group-level data.
**Research Example**: "Countries with more education spending have higher GDP, so spending on education makes individuals richer."
**Detection**: Are individual-level and group-level relationships the same?

### 20. Simpson's Paradox
**Description**: A trend present in subgroups reverses when groups are combined.
**Research Example**: Department A and B both show improving retention, but the university overall shows declining retention (due to shifting enrollment proportions).
**Detection**: Always check disaggregated data alongside aggregate.

### Reasoning Fallacies

### 21. Straw Man
**Description**: Misrepresenting an opponent's argument to make it easier to attack.
**Research Example**: Critic says "this method has limitations" → Author responds "my critic says the entire study is worthless."
**Detection**: Does the refutation address the actual criticism?

### 22. Moving the Goalposts
**Description**: Changing the criteria for success after seeing results.
**Research Example**: Defining "program success" as enrollment growth, then shifting to "student satisfaction" when enrollment drops.
**Detection**: Were success criteria pre-defined?

### 23. Slippery Slope
**Description**: Arguing that one action will inevitably lead to an extreme outcome.
**Research Example**: "If we allow flexible admission criteria, academic standards will collapse entirely."
**Detection**: Is each step in the chain actually probable?

### 24. Circular Reasoning (Begging the Question)
**Description**: The conclusion is assumed in the premise.
**Research Example**: "This university is excellent because it is highly ranked, and it is highly ranked because it is excellent."
**Detection**: Does the argument depend on the truth of what it's trying to prove?

### 25. No True Scotsman
**Description**: Redefining a category to exclude counterexamples.
**Research Example**: "All quality assurance systems improve outcomes." "But system X didn't." "Well, X wasn't a true quality assurance system."
**Detection**: Is the definition being modified to fit the claim?

### 26. Equivocation
**Description**: Using a term in two different senses within the same argument.
**Research Example**: "Quality" used sometimes to mean "standards compliance" and sometimes to mean "student satisfaction."
**Detection**: Is the key term defined consistently throughout?

### Statistical Fallacies

### 27. Base Rate Neglect
**Description**: Ignoring the base rate (overall probability) in favor of specific information.
**Research Example**: "This program has a 90% satisfaction rate" — but the base rate for all programs is 88%.
**Detection**: Always compare against relevant base rates.

### 28. Regression to the Mean
**Description**: Extreme performances naturally tend back toward average on subsequent measurements.
**Research Example**: "Our intervention improved scores for the lowest-performing students" — they may have improved anyway.
**Detection**: Was there a control group? Were initial measurements extreme?

### 29. Texas Sharpshooter
**Description**: Finding a pattern in random data by focusing on clusters and ignoring misses.
**Research Example**: Running 20 statistical tests and reporting only the 1 that was significant.
**Detection**: Were hypotheses pre-registered? Was multiple testing corrected for?

### 30. Gambler's Fallacy
**Description**: Believing past random events influence future random events.
**Research Example**: "This institution has declined for 5 years, so it's due for improvement."
**Detection**: Is there a causal mechanism for reversal, or is this just pattern-seeking?

### 31. McNamara Fallacy (Quantitative Bias)
**Description**: Making decisions based solely on quantitative metrics while ignoring qualitative factors.
**Research Example**: Ranking universities only by publication counts, ignoring teaching quality and community impact.
**Detection**: Are important but hard-to-measure factors being excluded?

### 32. Goodhart's Law
**Description**: "When a measure becomes a target, it ceases to be a good measure."
**Research Example**: Universities gaming rankings metrics instead of genuinely improving quality.
**Detection**: Has the metric become a target? Are there signs of metric manipulation?

## Quick Reference: Detection Questions

| Ask This | Detects |
|----------|---------|
| "Does B have other possible causes?" | Post hoc, false cause |
| "What about the failures?" | Survivorship bias |
| "Is this sample representative?" | Hasty generalization |
| "Were criteria defined before results?" | Moving goalposts, Texas sharpshooter |
| "Is the key term used consistently?" | Equivocation |
| "What's the base rate?" | Base rate neglect |
| "What evidence was left out?" | Cherry-picking, confirmation bias |
| "Is this the actual argument being made?" | Straw man |
| "Can we distinguish correlation from causation?" | Cum hoc, ecological fallacy |
| "Are individual and group levels being mixed?" | Ecological fallacy, Simpson's paradox |
writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deploymentStars: 323kUpdated Claude CodeCodexVerifiedInstall
xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release.Stars: 118kUpdated Claude CodeCodexVerifiedInstall