Whole Regrouper

by Goobee811

skill

|

Skill Details

Repository Files

8 files in this skill directory


name: whole-regrouper description: | Phân tích, gom nhóm, và ĐỒNG BỘ THÔNG MINH giữa Tổng Quan listing và actual group headers. v5.1.0: Intelligent Analysis with agent integration - phân tích thực sự cả hai với hỗ trợ từ specialized agents. version: 5.1.0 license: MIT allowed-tools:

  • Edit
  • Grep
  • Read
  • Bash
  • Task metadata: author: "Whole Project" category: "documentation" updated: "2026-01-02"

Whole Concept Regrouper & Reconciler v5.1

Intelligent Analysis with Agent Integration - Phân tích thực sự cả hai groupings với hỗ trợ từ specialized agents.


Core Philosophy

KHÔNG giả định grouping nào tốt hơn.

Mỗi CHỨC NĂNG có TWO representations:

  1. Tổng Quan listing - Overview với group names và concept counts
  2. Content headers - Actual ### headers với #### concepts bên dưới

Cả hai có thể có điểm mạnh riêng:

  • Tổng Quan có thể có grouping logic tốt hơn (coherent, mental model rõ)
  • Content có thể có chi tiết chính xác hơn (accurate to actual concepts)

→ Phân tích cả hai, quyết định có căn cứ.


Analysis Criteria

1. Coherence (Mạch lạc) - Weight: HIGH

  • Các concepts trong nhóm có chung chủ đề?
  • Có thể giải thích "đều về..." trong 1 câu?
  • Có concept "lạc lõng"?

2. Balance (Cân bằng) - Weight: MEDIUM

  • Per group: 3-8 concepts (ideal: 5-6)
  • Không có groups quá lớn (>10) hoặc quá nhỏ (<2)

3. Natural Thinking (Tự nhiên) - Weight: HIGH

  • Phù hợp mental model của người dùng?
  • Tên nhóm gợi nhớ ngay nội dung?

4. Accuracy (Chính xác) - Weight: MEDIUM

  • Tên nhóm mô tả chính xác nội dung?
  • Số concepts match?
  • Concept names chính xác?

Strategy Options

[A] Tổng Quan → Content
    Tổng Quan có grouping logic TỐT HƠN
    → Reorganize content để match Tổng Quan

[B] Content → Tổng Quan
    Content có chi tiết CHÍNH XÁC HƠN
    → Update Tổng Quan listing để match actual

[C] Full Regroup
    CẢ HAI ĐỀU CÓ VẤN ĐỀ
    → Cần phân tích lại từ đầu với /regroup

[H] Hybrid Merge
    MỖI BÊN CÓ ĐIỂM MẠNH RIÊNG
    → Lấy groups tốt nhất từ cả hai
    → Chỉ định: "Group 1,3 from Tổng Quan + Group 2,4 from Content"

[S] Skip - Already Synced
    Hai bên ĐÃ ĐỒNG BỘ
    → Không cần thay đổi

Decision Framework

Priority order khi conflict:

  1. Coherence > Balance

    • Grouping logic quan trọng hơn size
  2. Natural Thinking > Accuracy

    • User experience > technical correctness
  3. Khi tie → Consider Hybrid [H]

    • Lấy best of both worlds
  4. Khi cả hai < 3 sao → Full Regroup [C]

    • Cần làm lại từ đầu

Integration with Agents

When to Invoke Agents

Use Task tool to invoke specialized agents for deep analysis during regrouping:

// For semantic grouping analysis and duplicate detection
Task(subagent_type: 'whole-translator',
     prompt: 'Analyze semantic coherence and cultural grouping for CF[N] concepts')

// For cross-reference validation after regrouping
Task(subagent_type: 'whole-cross-reference',
     prompt: 'Validate and update cross-references after regrouping CF[N]')

// For structure validation before commit
Task(subagent_type: 'whole-content-validator',
     prompt: 'Validate regrouped structure and compliance for CF[N]')

When NOT to Use Agents

  • Simple reconciliation (strategy [S] - already synced) → Direct comparison
  • Balance checks (counting concepts) → Use Grep/scripts
  • Format validation → Use validation scripts in scripts/
  • Single group rename → Direct Edit

Agent Integration Guide

whole-translator

When to use: Semantic coherence analysis for grouping decisions Command: Task(subagent_type='whole-translator', prompt='Analyze semantic grouping coherence for CF[N]') Expected output: Semantic similarity analysis, cultural grouping recommendations

whole-cross-reference

When to use: Validate cross-references after major regrouping Command: Task(subagent_type='whole-cross-reference', prompt='Validate cross-references after CF[N] regroup') Expected output: Cross-reference validation report, orphaned links detection

whole-content-validator

When to use: Final validation before committing regrouped content Command: Task(subagent_type='whole-content-validator', prompt='Validate regrouped CF[N] structure') Expected output: Structure validation report, compliance check


Workflow

/reconcile [N]

Phase 1: LOCATE & READ
├─ Grep "## CHỨC NĂNG" → boundaries
└─ Read section content

Phase 2: PARSE BOTH
├─ A: Tổng Quan listing
└─ B: Content headers + concepts

Phase 3: ANALYZE
├─ Score each grouping on 4 criteria
├─ Compare winner per criterion
└─ Calculate overall score

Phase 4: RECOMMEND
├─ Reasoned recommendation [A/B/C/H]
├─ Explain trade-offs
└─ Ask for confirmation

Phase 5: EXECUTE
├─ Apply chosen strategy
├─ Validate changes
└─ Auto commit & push

/regroup [N]

Full regroup workflow khi cần phân tích lại từ đầu.


Scoring Output Format

╔═══════════════════════════════════════════════╗
║ ANALYSIS: CF[N] - [Function Name]             ║
╠═══════════════════════════════════════════════╣
║                                               ║
║ TỔNG QUAN: [M] groups                         ║
║ Coherence: ⭐⭐⭐⭐☆ | Balance: ⭐⭐⭐☆☆        ║
║ Natural: ⭐⭐⭐⭐⭐ | Accuracy: ⭐⭐⭐☆☆         ║
║                                               ║
║ CONTENT: [M] groups                           ║
║ Coherence: ⭐⭐⭐☆☆ | Balance: ⭐⭐⭐⭐☆        ║
║ Natural: ⭐⭐⭐☆☆ | Accuracy: ⭐⭐⭐⭐⭐         ║
║                                               ║
╠═══════════════════════════════════════════════╣
║ RECOMMENDATION: [A/B/C/H] - [Reasoning]       ║
╚═══════════════════════════════════════════════╝

Critical Rules

✅ MUST

  • Phân tích thực sự cả hai groupings (analyze both, don't assume)
  • Cho điểm có căn cứ (score with evidence)
  • Giải thích reasoning (explain trade-offs)
  • Preserve all content (only add, never subtract)
  • Use agents for deep semantic analysis when needed
  • Validate with whole-content-validator before commit
  • Use shared utilities from .claude/skills/shared
  • Update cross-references after major regrouping

❌ NEVER

  • Giả định một bên luôn đúng (assume one side is always right)
  • Skip analysis phase
  • Delete concepts or modify concept content
  • Commit without validation (use scripts or agents)
  • Ignore cross-reference integrity
  • Use agents for simple tasks (prefer scripts)

Commands

  • /reconcile [N] - Intelligent reconcile single CHỨC NĂNG
  • /reconcile - Auto-detect next pending
  • /regroup [N] - Full regroup when reconcile isn't enough

References

Load as needed:

  • references/grouping-principles.md - Criteria details
  • references/naming-guidelines.md - Naming standards
  • references/quality-checklist.md - Validation checklist

Version: 5.1.0 | Philosophy: Analyze first, decide with reasoning, validate with agents

Related Skills

Attack Tree Construction

Build comprehensive attack trees to visualize threat paths. Use when mapping attack scenarios, identifying defense gaps, or communicating security risks to stakeholders.

skill

Grafana Dashboards

Create and manage production Grafana dashboards for real-time visualization of system and application metrics. Use when building monitoring dashboards, visualizing metrics, or creating operational observability interfaces.

skill

Matplotlib

Foundational plotting library. Create line plots, scatter, bar, histograms, heatmaps, 3D, subplots, export PNG/PDF/SVG, for scientific visualization and publication figures.

skill

Scientific Visualization

Create publication figures with matplotlib/seaborn/plotly. Multi-panel layouts, error bars, significance markers, colorblind-safe, export PDF/EPS/TIFF, for journal-ready scientific plots.

skill

Seaborn

Statistical visualization. Scatter, box, violin, heatmaps, pair plots, regression, correlation matrices, KDE, faceted plots, for exploratory analysis and publication figures.

skill

Shap

Model interpretability and explainability using SHAP (SHapley Additive exPlanations). Use this skill when explaining machine learning model predictions, computing feature importance, generating SHAP plots (waterfall, beeswarm, bar, scatter, force, heatmap), debugging models, analyzing model bias or fairness, comparing models, or implementing explainable AI. Works with tree-based models (XGBoost, LightGBM, Random Forest), deep learning (TensorFlow, PyTorch), linear models, and any black-box model

skill

Pydeseq2

Differential gene expression analysis (Python DESeq2). Identify DE genes from bulk RNA-seq counts, Wald tests, FDR correction, volcano/MA plots, for RNA-seq analysis.

skill

Query Writing

For writing and executing SQL queries - from simple single-table queries to complex multi-table JOINs and aggregations

skill

Pydeseq2

Differential gene expression analysis (Python DESeq2). Identify DE genes from bulk RNA-seq counts, Wald tests, FDR correction, volcano/MA plots, for RNA-seq analysis.

skill

Scientific Visualization

Meta-skill for publication-ready figures. Use when creating journal submission figures requiring multi-panel layouts, significance annotations, error bars, colorblind-safe palettes, and specific journal formatting (Nature, Science, Cell). Orchestrates matplotlib/seaborn/plotly with publication styles. For quick exploration use seaborn or plotly directly.

skill

Skill Information

Category:Skill
License:MIT
Version:5.1.0
Last Updated:1/3/2026