Claude Code+Obsidian+技能图谱：构建本地研究引擎

# Research Skill Graph — Command Center## 1. MissionDeep research system that takes ONE question and produces a multi-angle analysis no single Google search or ChatGPT prompt could ever match.Instead of 50 open tabs and scattered notes, this system forces structured thinking through 6 research lenses, each one rethinking the question from a fundamentally different angle.Research Question: [PASTE YOUR QUESTION HERE]Scope: [DEFINE BOUNDARIES — what's in, what's out]Time Horizon: [how far back and forward are we looking?]Output Goal: [what decision does this research inform?]## Prior Research (optional — for compound mode)Check [[research-log]] for previous research that may connect to this question.Relevant prior projects: [link any related projects from research-log.md, or leave empty for clean slate]## 2. Node MapEvery node below is a knowledge file. Read the relevant ones before executing any task. The [[wikilinks]] are clickable — follow them.### Methodology- [[research-frameworks]] — how to approach different types of questions. start here to pick the right research structure- [[source-evaluation]] — criteria for judging if a source is worth trusting. tier system from primary data to random blog posts- [[synthesis-rules]] — how to combine findings across lenses without losing nuance. the hardest part of research- [[contradiction-protocol]] — what to do when sources disagree. this is where the real insights hide### Lenses (the core engine)- [[technical]] — how does it work mechanically? what do the numbers actually say? strip away narrative, look at data- [[economic]] — follow the money. who pays, who profits, what markets move, what incentives drive behavior- [[historical]] — what patterns repeat? what's been tried before? what context does everyone forget?- [[geopolitical]] — which countries, which power dynamics, which alliances and conflicts shape this?- [[contrarian]] — what if the consensus is wrong? who benefits from the current narrative? what's nobody saying?- [[first-principles]] — forget everything you think you know. rebuild from fundamental truths only### Outputs- [[executive-summary]] — the final synthesis. 500 words max. what did we learn, what does it mean, what's still unknown- [[deep-dive]] — the full analysis organized by lens, with cross-references and contradictions highlighted- [[key-players]] — people, organizations, countries that matter most on this topic- [[open-questions]] — what we STILL don't know after research. often more valuable than what we found### Knowledge Base- [[concepts]] — key terms, definitions, mental models relevant to the research- [[data-points]] — hard numbers, statistics, metrics collected during research. always with source attribution### Sources- [[source-template]] — template for processing raw sources into structured notes## 3. Execution InstructionsWhen given a research question:1. Read this file completely. understand the scope and goal2. Read [[research-frameworks]] to pick the right approach for this type of question3. Read [[source-evaluation]] so you know what counts as good evidence4. For EACH lens in order (technical → economic → historical → geopolitical → contrarian → first-principles):   a. Read the lens file for its specific angle and questions   b. Research the topic THROUGH that lens only   c. Record findings, sources, and confidence level   d. Note any contradictions with previous lenses5. Read [[contradiction-protocol]] — resolve or document disagreements between lenses6. Read [[synthesis-rules]] — combine everything7. Produce all 4 output files: [[executive-summary]], [[deep-dive]], [[key-players]], [[open-questions]]8. Update [[concepts]] and [[data-points]] with everything learnedCRITICAL RULE: each lens must RETHINK the question, not just add more information. the technical lens and the contrarian lens should feel like they were written by two different researchers who disagree with each other. that tension is where insight lives.

# Research FrameworksHow to approach different types of research questions. Read this BEFORE starting any research to pick the right structure.## Framework Selection### Type 1: "Is X true?" (Verification)- Start with [[technical]] lens to establish what the data actually says- Then [[contrarian]] to stress-test the claim- Then [[historical]] for precedent- Best for: fact-checking, debunking, validating assumptions- Example: "Is nuclear fusion viable by 2035?"### Type 2: "Why is X happening?" (Causal Analysis)- Start with [[historical]] to trace the roots- Then [[economic]] to find incentive structures- Then [[technical]] for mechanism- Then [[geopolitical]] for systemic forces- Then [[contrarian]] to challenge your causal chain- Best for: understanding trends, explaining phenomena- Example: "Why are birth rates collapsing globally?"### Type 3: "What happens if X?" (Scenario Planning)- Start with [[first-principles]] to establish base assumptions- Then [[technical]] for constraints- Then [[economic]] for incentives- Then [[geopolitical]] for power dynamics- Best for: forecasting, strategy, decision-making- Example: "What happens to Europe if birth rates stay below replacement for 30 years?"### Type 4: "What should I do about X?" (Decision Support)- Start with [[executive-summary]] of existing knowledge- Then run all 6 lenses in parallel- Rank options by lens agreement (if 5/6 lenses point the same way, high confidence)- Best for: investment decisions, policy, strategy## Research Depth Levels### Level 1: Quick Scan (30 min)- 3 lenses maximum- Top 5 sources only- Goal: directional understanding, not certainty### Level 2: Standard Research (2-3 hours)- All 6 lenses- 15-25 sources- Cross-reference findings between lenses- Goal: informed opinion backed by evidence### Level 3: Deep Dive (1-2 days)- All 6 lenses with sub-questions- 50+ sources including primary data- Interview/expert source integration- Contradiction resolution required- Goal: publishable analysis, decision-grade intelligence## Source Collection StrategyFor each lens:1. Start with the BEST single source (see [[source-evaluation]])2. Find the source that DISAGREES most with #13. Find primary data that lets you judge between them4. Record everything in [[data-points]] with attribution

# Source EvaluationHow to judge if a source is worth trusting. Apply this to EVERY source before using it in analysis.## Source Tier System### Tier 1: Primary Data (highest trust)- Raw datasets (UN, World Bank, national statistics offices)- Peer-reviewed studies with methodology visible- Financial filings, government records- Direct measurements and observations- USE FOR: [[data-points]], hard claims, base assumptions### Tier 2: Expert Analysis- Reports from domain-specific research institutions- Books by recognized authorities in the field- Long-form investigative journalism with cited sources- Conference papers and working papers- USE FOR: interpretation, causal claims, framework building### Tier 3: Informed Commentary- Expert blog posts and newsletters- Quality podcasts with domain experts- Think tank reports (check funding sources)- Industry publications- USE FOR: angles you hadn't considered, hypothesis generation### Tier 4: General Media- Major news outlets- Wikipedia (good for overview, never for final claims)- Popular science writing- USE FOR: initial orientation only. always verify upstream### Tier 5: Social/Anecdotal (lowest trust)- Twitter threads, Reddit posts- Personal anecdotes- Viral content- USE FOR: signal detection only. "people are talking about X" ≠ "X is true"## Red Flags (downgrade any source by 1 tier if present)- No cited sources or methodology- Author has financial incentive in the conclusion- Published by organization with known agenda on the topic- Cherry-picked time frames or geographies- Conflates correlation with causation- Uses emotional language instead of evidence## Evaluation ChecklistFor every key claim in your research, ask:1. What tier is this source?2. Can I find the same claim in a Tier 1 or Tier 2 source?3. Who funded this research or publication?4. What would the author lose if they were wrong?5. Is this the BEST available evidence, or just the first I found?See [[contradiction-protocol]] when two credible sources disagree.

# Synthesis RulesHow to combine findings across lenses without losing nuance. This is the hardest part of research — most people skip it and just stack facts.## The Synthesis Process### Step 1: Lens SummaryAfter completing all 6 lenses, write a ONE paragraph summary per lens:- What is this lens's main finding?- Confidence level: high / medium / low- What surprised you from this angle?### Step 2: Agreement MapIdentify where lenses AGREE:- If 4+ lenses point the same direction → high confidence finding- If 3 lenses agree → moderate confidence, worth stating with caveats- If only 1-2 lenses support a claim → hypothesis only, flag as uncertain### Step 3: Tension MapIdentify where lenses DISAGREE:- [[technical]] says X but [[economic]] says Y → this tension IS the insight- Don't resolve by picking a winner. document both positions- Ask: "under what conditions is each lens correct?"- See [[contradiction-protocol]] for resolution framework### Step 4: Second-Order InsightsThe best findings come from COMBINING lenses:- "The technical data shows declining fertility, but the economic lens reveals that financial incentives haven't reversed it anywhere — which means the [[first-principles]] lens's argument about cultural shifts might be the dominant factor"- These cross-lens insights are what make this system more powerful than reading 50 articles### Step 5: Confidence CalibrationFor each major finding, state:- CLAIM: what you believe is true- EVIDENCE: strongest supporting data (from [[data-points]])- CONFIDENCE: high / medium / low- WHAT WOULD CHANGE MY MIND: the specific evidence that would reverse this conclusion## Output Rules- Never present a single-lens finding as a conclusion- Always show the tension between lenses- Separate "what the data shows" from "what I interpret"- [[open-questions]] is as important as [[executive-summary]]- Prefer "it seems likely that X because Y and Z, though A complicates this" over "X is true"## Anti-Patterns (things that kill good research)- Confirmation bias: only searching for evidence that supports your initial hunch- Narrative fallacy: making a clean story out of messy, contradictory data- Recency bias: overweighting the latest article over 30 years of data- Authority bias: believing something because an impressive person said it- Anchoring: letting the first number you found define your mental range

# Contradiction ProtocolWhat to do when sources or lenses disagree. This is where the real insights hide — contradictions are features, not bugs.## When Two Sources Disagree### Step 1: Check the basics- Are they talking about the same thing? (different geographies, time periods, definitions)- Are they using the same data? (same dataset, different interpretation?)- Is one source a higher tier than the other? (see [[source-evaluation]])### Step 2: Find the root of disagreementUsually one of:- Different data: they're looking at different datasets → find out which is more complete/recent- Different interpretation: same data, different conclusions → examine their reasoning chain- Different scope: one is global, one is country-specific → both might be right in their context- Different timeframe: short-term vs long-term trends can look opposite- Different incentives: one has reason to spin the data → check funding, affiliations### Step 3: Document, don't resolveIn [[deep-dive]], write it as:"Source A argues [X] based on [data]. Source B argues [Y] based on [data]. The disagreement likely stems from [root cause]. Under conditions [C1], A is probably right. Under conditions [C2], B is probably right."### Step 4: Upgrade to [[open-questions]] if unresolvableIf after honest effort you can't determine who's right:- This IS a finding. "We don't know whether X or Y" is valuable intelligence- Add it to [[open-questions]] with the specific evidence needed to resolve it- This often becomes the most interesting part of the research## When Two Lenses DisagreeThis is expected and healthy. Different lenses SHOULD produce tension.## Confidence Adjustment- 2 lenses agree, 1 disagrees → investigate the disagreeing lens deeper. it might see something the others miss- All lenses agree → be suspicious. you might have confirmation bias. re-read [[contrarian]]- No lenses agree → your research question might be too broad. narrow the scope in [[index]]

# Lens: TechnicalStrip away opinions and narratives. What do the numbers actually say? What mechanisms are at work?## Core Questions1. What does the DATA show? (not what people say about the data)2. What are the measurable inputs and outputs?3. What mechanisms drive the phenomenon?4. What are the hard constraints (physical, biological, mathematical)?5. What metrics matter most and how are they measured?6. Where is the data incomplete or poorly measured?## How to Research Through This Lens- Start with Tier 1 sources ONLY (see [[source-evaluation]]): raw datasets, peer-reviewed studies- Look for: time series, geographic comparisons, demographic breakdowns- Quantify everything. replace "declining" with "declined by X% between Y and Z"- Identify measurement problems: how is this data collected? what's excluded?- Find the base rates. what's "normal"? what's the historical range?## Output FormatFor each finding:- METRIC: [what you measured]- VALUE: [the number]- SOURCE: [Tier 1 or 2 source with date]- TREND: [direction and rate of change]- CAVEAT: [measurement limitations]Record all hard numbers in [[data-points]].## VoiceClinical. precise. no emotional language. "fertility rate dropped from 2.1 to 1.6" not "fertility is collapsing." let the numbers speak.## Connects To- [[source-evaluation]] — only high-tier sources for this lens- [[data-points]] — all numbers go here- [[economic]] — technical data often explains economic outcomes- [[first-principles]] — technical constraints define what's possible

# Source Template## Source: [Title]- Author: [who]- Date: [when published]- URL: [link]- Tier: [1-5, from [[source-evaluation]]]- Lens: [which lens found this source]## Key Claims1. [claim with page/section reference]2. [claim]3. [claim]## Data Extracted[any numbers → also add to [[data-points]]]## Methodology Notes[how did they arrive at their conclusions? any red flags?]## Connections[what other sources or concepts does this relate to? use [[wikilinks]]]## My Assessment[is this trustworthy? what's strong, what's weak?]