Research Skill
Systematic research query expansion and completion assessment. Transforms basic questions into comprehensive search strategies.
Diagnostic States
R1: No Analysis
Symptoms: Jumping straight to searching without analyzing the topic.
Test: Can you articulate stakeholders, temporal scope, and domain mapping?
Intervention: Run Phase 0 Analysis Template before generating queries.
R1.5: No Vocabulary Map
Symptoms: Using outsider/introductory terminology. Finding only surface-level material.
Test: Have you identified expert vs. outsider terms? Terms across domains?
Intervention: Build vocabulary map. Hunt for "also known as," "technically called" in early sources.
R2: Single-Perspective Search
Symptoms: All queries support one viewpoint. Missing counterarguments.
Test: Have you explicitly searched for opposing perspectives?
Intervention: Generate competing perspectives queries. Search for strongest counterargument.
R3: Domain Blindness
Symptoms: Searching only in familiar field. Missing cross-disciplinary insights.
Test: Have you mapped terminology variants across fields?
Intervention: Identify what adjacent fields call this topic. Search in at least 2 domains. Update vocabulary map.
R4: Recency Bias
Symptoms: Only recent sources. Missing historical context.
Test: Can you explain when this topic emerged and how it evolved?
Intervention: Add historical context queries. Find seminal works.
R5: Breadth Without Depth
Symptoms: Many tabs, no synthesis. Can't explain core concepts.
Test: Can you define key terms in your own words?
Intervention: Apply 3-source rule per perspective. Summarize before searching more.
R6: Completion Uncertainty
Symptoms: Unsure whether to continue or stop. Research expanding indefinitely.
Test: Can you answer the tiered completion criteria?
Intervention: Run completion checklist. Look for diminishing returns signals.
R7: Research Complete
Symptoms: Can explain topic, identify uncertainties, and take action.
Indicators: Circular references, repetitive findings, sufficient for purpose.
R8: No Persistence
Symptoms: Starting from scratch each session. Re-discovering same vocabulary.
Test: Did you check for prior research before starting? Are you storing findings?
Intervention: Store vocabulary map, sources, digested notes, and gaps for future use.
R9: Scope Mismatch
Symptoms: Over-researching trivial questions. Under-researching critical decisions.
Test: Is research depth proportional to decision stakes?
Intervention: Apply scope calibration. Match confidence level to decision reversibility and stakes.
R10: No Confidence Signaling
Symptoms: Hedging language everywhere. Reader can't tell what's certain vs. speculative.
Test: Can reader distinguish established facts from speculation?
Intervention: Use explicit confidence markers. State source quality and consensus status.
Phase 0: Analysis Template
Before searching, structure your topic:
markdown
# Research Analysis: [Topic]
## Core Concepts
- **Primary terms:** [Key terms requiring definition]
- **Terminology variants:** [Synonyms, jargon, historical terms]
- **Ambiguous terms:** [Terms with multiple meanings]
## Stakeholders
- **Primary actors:** [Who is directly involved?]
- **Affected groups:** [Who bears consequences?]
- **Opposing interests:** [Who benefits from different outcomes?]
## Temporal Scope
- **Historical origins:** [When did this begin?]
- **Key transitions:** [What changed and when?]
- **Current state:** [What's happening now?]
## Domains
- **Primary field:** [Main discipline]
- **Adjacent fields:** [Related disciplines]
## Controversies
- **Active debates:** [What's contested?]
- **Competing frameworks:** [Different ways of understanding]
Query Types
- Foundational: "term definition AND field review"
- Historical: "topic history development [date range]"
- Current: "topic current trends [recent years]"
- Competing: "topic debate AND (perspective1 OR perspective2)"
- Evidence: "topic impact measurement study data"
Completion Criteria
Minimum Viable (Quick Decisions)
Working Knowledge (Most Decisions)
Deep Expertise (High-Stakes)
Diminishing Returns Signals
Stop when:
- New sources cite same foundational works (circular)
- New searches return familiar content (repetitive)
- Each hour adds less than previous (marginal)
- Can make decision or take action (sufficient)
Anti-Patterns
| Pattern | Symptom | Fix |
|---|
| Confirmation Trap | Searching to confirm, not learn | Search for strongest counterargument |
| Authority Fallacy | Accepting claims by source prestige | Evaluate evidence, not source |
| Recency Trap | Only recent sources | Explicitly search historical periods |
| Breadth Trap | 50 tabs, none read | 3-source rule, summarize before continuing |
| Single-Source | Wikipedia as final answer | Require 3 independent sources |
| Jargon Blind Spot | Missing other fields' terminology | Map variants, search multiple domains |
| Infinite Rabbit Hole | Lost original purpose | Write decision/action anchor, return to it |
Vocabulary Mapping
Primary research deliverable. Vocabulary determines search space and LLM semantic activation.
Why It Matters
- Expert terms → expert material. Outsider terms → introductory material.
- Precise vocabulary activates richer LLM semantic space.
- Cross-domain terms bridge bodies of work that use different names.
Vocabulary Map Template
markdown
## Core Terms
|------|--------|-------------|
| [expert term] | [field] | Expert |
| [outsider term] | General | Introductory |
## Cross-Domain Synonyms
|---------|-----------------|
| [concept] | Field A: [term], Field B: [term] |
## Depth Indicators
|-------|-------|-------------------|
| Introductory | [terms] | Overviews, explainers |
| Expert | [terms] | Research, nuanced analysis |
Discovery Process
- Note which terms feel like outsider language
- In early sources, watch for "also known as," "technically called"
- Map terms across domains
- Test different terms in searches, note what surfaces
Research Persistence
Store both sources AND digested results. Don't start from scratch.
What to Store
| Layer | Contents |
|---|
| Vocabulary Map | Terms, domains, depth levels |
| Sources | PDFs, saved pages, bookmarks |
| Digested Notes | Summaries, key quotes, synthesis |
| Query Log | Searches that worked/failed |
| Gaps | What remains unknown |
Before Starting
Check for prior research. Load vocabulary map. Start where you left off.
Single-Shot Research
When research runs without follow-up questions (agent execution, time-boxed queries):
Scope Calibration
| Decision Type | Confidence Needed | Research Depth |
|---|
| Reversible, low-stakes | 60-70% | Quick scan (minutes) |
| Reversible, moderate | 75-85% | Working knowledge (1-2 hours) |
| Irreversible, moderate | 85-90% | Solid grounding (half day) |
| Irreversible, high | 90-95% | Deep expertise (days) |
Question Pattern → Strategy
| Pattern | Strategy |
|---|
| "What is X?" | 2-3 authoritative sources, establish consensus |
| "Should I X?" | Pros/cons, alternatives, conditions for each |
| "Is X true?" | Primary sources, counter-evidence, consensus check |
| "How do I X?" | Step-by-step, prerequisites, common pitfalls |
Source Type Selection
| Source | Best For |
|---|
| Wikipedia/Encyclopedias | Orientation, terminology, citation hunting |
| Academic papers | Mechanism, causation, methodology |
| Practitioner content | How things actually work, edge cases |
| Official docs | Technical specs, policy, procedures |
Synthesis Template
markdown
## Summary
[Direct answer to question]
## Confidence Level
[High/Medium/Low] - [Justification]
## Key Findings
1. [Finding with source type]
## Caveats
- [What wasn't consulted]
- [What assumptions were made]
## For Deeper Investigation
[What would increase confidence]
Confidence Markers
| Level | Phrases |
|---|
| Established | "X is...", "X works by..." |
| Strong evidence | "Evidence strongly suggests..." |
| Moderate evidence | "Most sources report..." |
| Limited evidence | "One study found..." |
| Unknown | "No reliable information found..." |
Single-Shot Checklist
Health Check Questions
During research, ask:
- Am I searching to learn or to confirm?
- What's the strongest argument against my current view?
- Have I looked outside my familiar domains?
- Can I summarize what I've learned so far?
- Is this still serving my original purpose?
- Am I using expert or outsider vocabulary?
- Have I stored what I've learned for future use?
- Is my depth proportional to the stakes?
- Am I signaling confidence explicitly?
Integration Points
| Skill | Connection |
|---|
| doppelganger | Research informs decisions; apply /truth-check to findings |
| context-networks | Store research findings in appropriate network node |
| boundary-critique | Apply to advice and recommendations encountered |
Output Persistence
This skill writes primary output to files so work persists across sessions.
Output Discovery
Before doing any other work:
- Check for in the project
- If found, look for this skill's entry
- If not found or no entry for this skill, ask the user first:
- "Where should I save output from this research session?"
- Suggest: or a sensible location for this project
- Store the user's preference:
- In if context network exists
- In at project root otherwise
Primary Output
For this skill, persist:
- Vocabulary map - terms, domains, depth levels, cross-domain synonyms
- Phase 0 analysis - core concepts, stakeholders, temporal scope, domains
- Synthesis document - summary, confidence levels, key findings, caveats
- Source assessment - sources consulted with quality notes
- Gaps identified - what remains unknown, next steps
Conversation vs. File
| Goes to File | Stays in Conversation |
|---|
| Vocabulary map | Discussion of term discovery |
| Synthesis document | Query refinement iterations |
| Source list with assessments | Real-time source evaluation |
| Gap analysis | Clarifying questions |
| Confidence-marked findings | Follow-up investigations |
File Naming
Pattern:
{topic}-research-{date}.md
Example:
competency-frameworks-research-2025-01-15.md
Relationship to Research Persistence Section
The "Research Persistence" section above describes WHAT to store. This section operationalizes WHERE and HOW - ensuring the skill checks for configured locations, asks the user when needed, and writes output consistently.
Source Framework
Derived from:
frameworks/research/research-framework.md