AgentSkillsCN

Practitioner Research Skill

在 AI 辅助开发的背景下,探索纵向案例研究的方法论,以循证为基础开展文献整理,并产出符合出版标准的学术写作成果。

SKILL.md
--- frontmatter
name: "Practitioner Research Skill"
description: "Practitioner methodology for longitudinal case study research, evidence-based documentation, and publication-ready academic writing in AI-assisted development."

Practitioner Research Skill

Evidence-based research methodology for longitudinal case studies with publication-ready output.

This skill captures the practitioner research approach: practitioner-first, evidence-grounded, visually rich methodology papers that bridge academic rigor with real-world applicability.

Core Philosophy

PrincipleDescription
Ship → Document → PromoteResearch what you've done, not what you plan. Theory follows practice.
Visual LearningDiagrams are primary evidence. Every architecture decision gets a Mermaid.
Structured AbstractsBackground, Objective, Method, Results, Contributions, Significance
Dual AudiencePart I: Universal Framework (any reader). Part II: Your Journey (case study)
Living DocumentsGit-tracked, evolving with code. DRY—no duplicate content.

Research Workflow

Phase 1: Practice First (3-18 months)

mermaid
%%{init: {'theme': 'base', 'themeVariables': { 'edgeLabelBackground': '#f6f8fa'}}}%%
flowchart LR
    P1["🚀 Ship Projects"] --> P2["📝 Capture Insights"] --> P3["🔍 Notice Patterns"]
    P3 --> P4["💡 Publication Potential"]
    
    style P1 fill:#e8f5e9,stroke:#2e7d32
    style P2 fill:#e3f2fd,stroke:#1565c0
    style P3 fill:#fff3e0,stroke:#ef6c00
    style P4 fill:#f3e5f5,stroke:#7b1fa2
  • Work extensively in the domain (62 projects, 148 insights)
  • Let patterns emerge through doing
  • Global knowledge captures cross-project learnings
  • Only write the paper AFTER the methodology is proven

Phase 2: Document Structure

The paper architecture follows this template:

  1. Abstract (Structured)

    • Background: Problem context and gap
    • Objective: Research question
    • Method: What you did
    • Results: Key findings (quantified)
    • Contributions: Novel concepts introduced
    • Significance: Why it matters
  2. Foundational Insight — Origin story, what sparked the research

  3. Part I: Universal Framework — Generalizable principles anyone can use

  4. Part II: Practitioner's Journey — Your specific implementation with evidence

  5. Appendices — Raw data, project inventory, literature context

Phase 3: Evidence Collection

Evidence TypeExampleHow to Collect
Quantitative metrics4-6× acceleration, 148 insightsTrack actual vs estimated time
Project inventory62 projects with success ratingsTable with pattern codes
Root cause analysisAnti-patterns and solutionsChronicle incidents as they happen
Literature contextPrior art vs novel contributionsSplit table format
Portfolio evidencePublic GitHub reposLinks with descriptions

Document Templates

Structured Abstract Template

markdown
**Background.** [Problem context. What limitation exists?]

**Objective.** [Research question. What are you trying to answer?]

**Method.** [What you did. Framework name, duration, scale.]

**Results.** [Key findings with numbers. Quantify everything.]

**Contributions.** [Novel concepts introduced. Name them explicitly.]

**Significance.** [Why it matters. What gap does this fill in literature?]

Prior Art Table Template

ConceptPrior ArtOur Extension
{Concept name}{Citation}: {key finding}{How we extended or applied it}

Competitive Analysis Table Template

CapabilityCompetitor ACompetitor BThis Work
{Feature}⭐⭐⭐⭐⭐⭐⭐⭐

Legend: ⭐ = basic, ⭐⭐ = good, ⭐⭐⭐ = excellent, ❌ = missing

Project Inventory Table Template

ProjectSkillsSuccessPatternNotes
{Name}{count}⭐⭐⭐⭐⭐🔄{Brief note}

Pattern Codes:

  • 🚀 Quick win — completed in scope
  • 🔄 Ongoing — living system
  • 🎯 Promotion candidate — should earn skills
  • ⚠️ Scope issue — unrealistic goals
  • 🚧 Blocked — external dependencies
  • 📦 Archive candidate — inactive

Visual Research Standards

Mermaid Color Palette

javascript
%%{init: {'theme': 'base', 'themeVariables': { 'edgeLabelBackground': '#f6f8fa'}}}%%
ColorHexUsage
Info Blue#e3f2fd / #1565c0Neutral information, starting states
Success Green#e8f5e9 / #2e7d32Positive outcomes, solutions
Warning Orange#fff3e0 / #ef6c00Attention needed, transitions
Special Purple#f3e5f5 / #7b1fa2Unique concepts, outcomes
Danger Red#ffebee / #c62828Anti-patterns, problems

When to Use Which Diagram

Diagram TypeBest For
flowchartProcesses, workflows, architecture
quadrantChartPositioning, trade-offs
xychart-betaBar charts with metrics
pieDistribution, composition
mindmapTaxonomy, categorization
block-betaLayered systems, defense in depth

Citation Standards

APA 7 Quick Reference

markdown
## Single Author
Author, A. A. (Year). Title of work. *Journal Name, Volume*(Issue), pages. https://doi.org/xxx

## Multiple Authors (3+)
Author, A. A., Author, B. B., & Author, C. C. (Year). Title. *Journal*, pages.

## Web Sources
Author. (Year, Month Day). Title. *Site Name*. https://url

Fact-Checking Protocol

Before publication, verify:

  1. Author names — Correct spelling and order
  2. Dates — Publication year, access dates
  3. Statistics — Exact figures, not approximations
  4. DOIs — All links resolve
  5. Referenced vs cited — Everything cited appears in References

Prior Art vs Novel Contributions

Always distinguish:

TypeTreatment
Prior Art"builds on established research (Citation)"
Novel Contribution"We introduce..." / "This paper presents..."
Extension"We extend [Concept] by..."

Research Question Development

The Core Question Pattern

How can [method/technology] help [users] achieve [outcome] over [timeframe], not just [immediate benefit]?

Example:

How can AI help developers get smarter over time, not just faster today?

FINER for Practitioner Research

CriterionPractitioner Lens
FeasibleCan I do this with projects I'm already working on?
InterestingWill practitioners care? Will academics cite?
NovelWhat do I know from practice that literature doesn't cover?
EthicalAm I representing my own work fairly?
RelevantDoes this help someone else do better work?

Anti-Patterns to Avoid

Anti-PatternProblemSolution
Theory before practiceSkills written before doingShip first, document after
Overclaiming"Revolutionary" / "First ever"Precise language, acknowledge prior art
Buried evidenceData in prose, not tablesTables and diagrams primary
Stale citationsAll references >5 years oldMix foundational + contemporary
Self-selection bias"All my projects succeeded"Include failures, analyze patterns

Publication Pipeline

mermaid
%%{init: {'theme': 'base', 'themeVariables': { 'edgeLabelBackground': '#f6f8fa'}}}%%
flowchart TB
    subgraph Practice["Phase 1: Practice"]
        A["Projects"] --> B["Insights"] --> C["Patterns"]
    end
    
    subgraph Document["Phase 2: Document"]
        D["Draft"] --> E["Diagrams"] --> F["Evidence Tables"]
    end
    
    subgraph Polish["Phase 3: Polish"]
        G["Literature Review"] --> H["Fact-Check"] --> I["Academic Structure"]
    end
    
    subgraph Publish["Phase 4: Publish"]
        J["Peer Review"] --> K["Revise"] --> L["Submit"]
    end
    
    Practice --> Document --> Polish --> Publish
    
    style Practice fill:#e3f2fd,stroke:#1565c0
    style Document fill:#fff3e0,stroke:#ef6c00
    style Polish fill:#f3e5f5,stroke:#7b1fa2
    style Publish fill:#e8f5e9,stroke:#2e7d32

Session Types for Research

Session TypePurposeOutput
Evidence gatheringCompile metrics, project dataTables, inventories
Diagram creationVisualize conceptsMermaid diagrams
Literature searchFind and cite prior artReference list
Fact-checkingVerify claims and citationsCorrections
Academic polishStructure, abstract, conclusionPublication-ready sections

Metrics That Matter

MetricTargetWhy
Projects in evidence20+Demonstrates breadth
Quantified claimsAll major claimsAcademic rigor
Diagrams10-20 per paperVisual learning
References15-25Scholarly credibility
Novel contributions3-5 named conceptsPublishable value

Integration with Alex Architecture

This research method leverages:

  • Global Knowledge — Cross-project insights become evidence
  • Meditation Protocols — Consolidation captures learnable patterns
  • Skills from Doing — Only document proven methodologies
  • Synaptic Connections — Related concepts linked explicitly

Troubleshooting

"My paper feels like a blog post"

Problem: Informal tone, missing academic structure

Solution:

  1. Add structured abstract
  2. Create Prior Art table
  3. Add References section with APA 7
  4. Split Part I (Universal) / Part II (Specific)

"I don't have enough evidence"

Problem: Thin case study, few projects

Solution:

  1. Wait until you have 20+ projects
  2. Include failures and anti-patterns
  3. Add quantified metrics (time, count, percentage)

"My citations are all self-citations"

Problem: Looks like self-promotion

Solution:

  1. Add literature context section
  2. Show prior art for each concept
  3. Distinguish "builds on" vs "introduces"

This skill was earned through writing AI-ASSISTED-DEVELOPMENT-METHODOLOGY.md — 1500+ lines, 22+ references, 17+ diagrams, fact-checked and publication-ready.