Personal Knowledge Growth Analyzer

English | 简体中文

Personal Knowledge Growth Analyzer is a local knowledge analysis tool built on Claude Skills that helps you observe how you accumulate, record, and organize knowledge over time.

The Skill scans the materials you produce over time and helps you review your knowledge inputs and cognitive evolution through stage structure, recording patterns, and topic shifts.
It turns scattered folder activity into a traceable timeline, emphasizing the process of knowledge accumulation so you can see your learning rhythm and changes over time.

Positioning

Personal Knowledge Growth Analyzer

A local analysis tool that runs on your file system
Directly analyzes your existing folder structure and text materials
Uses stages as the unit to observe changes in knowledge activity
Combines structured statistics with AI semantic interpretation for traceable snapshots
Focuses on stage-level change and structure: “traceable, comparable, correctable”

What this tool intentionally does not do

Build a personal knowledge base
Provide Q&A or semantic search
Evaluate ability levels or personality traits
Produce “who you are” type conclusions

Results are always observational, not definitive.

Core Principles

Focus on change, not final judgments
Emphasize stage analysis over one-time summaries
Outputs are traceable, comparable, and correctable
Respect existing habits; no forced migration or restructuring

Typical Use Cases

1. Reviewing “wild” local knowledge accumulation

Many people’s knowledge lives outside formal note systems and instead in local folders, such as:

PDFs downloaded over time
Markdown / TXT notes
Word / RTF summaries
Mixed subfolders and topic files

These materials often reflect a phase of learning or a project but lack a coherent review perspective.

This Skill makes them analyzable without forcing you to reorganize or migrate.

2. Stage tracking for project-based learning or research

When you study a topic or run a project over time, materials evolve by stage:

Early: collection and quick notes
Middle: structured review and method summaries
Late: outputs, retrospectives, or pivots

This tool helps you observe whether those transitions are clear and consistent.

3. Stage review rather than one-time summaries

If you keep long-term records on a theme but struggle to compare “past thinking” with “current direction,” this tool supports stage review. Examples:

Career change prep: information collection → capability assessment → decision trade-offs
Content creation: exploration → methodology formation → stable production
Learning path: fragmented inputs → systematic synthesis → structured output

The value is seeing differences and turning points between stages, not producing a single conclusion.

Who It’s For

People with ongoing writing or note-taking habits
Knowledge workers, researchers, humanities/social science users
Those who store long-term materials in local folders
Anyone who wants to observe changes in thinking and recording style

If your records are sparse or highly fragmented, outputs may be thin.

How It Works

Stage Analysis

Stages are periods with higher input density and structural characteristics. Each run analyzes at most one recent stage and generates a snapshot.

First Run vs. Subsequent Runs

First run
- Looks back ~6 months
- Establishes an initial stage baseline
- Avoids performance issues with large history
Subsequent runs
- Analyze only new materials
- Keep all historical outputs
- Observe change rather than re-scan

Baseline (Long-term portrait)

Baseline is a low-frequency, long-term reference layer used to accumulate stable signals across stages.

Baseline is not a prerequisite
Stage analysis runs fully even if Baseline is unconfirmed
Baseline only affects the semantic weight of long-term conclusions

Baseline is unconfirmed by default and becomes a long-term reference only after explicit confirmation.

How to Run

Using Skills in Claude Code:

BASH

mkdir -p ~/.config/claude-code/skills/
cp -r personal-knowledge-growth-analyzer ~/.config/claude-code/skills/

Run these commands from the path that contains the personal-knowledge-growth-analyzer directory.

Verify skill metadata:

BASH

head ~/.config/claude-code/skills/personal-knowledge-growth-analyzer/SKILL.md

Start Claude Code:

BASH

claude

The Skill loads when Claude Code starts and can be invoked by explicit request or by the agent according to SKILL.md.

Run modes:

First run: look back ~6 months and generate a stage snapshot
Subsequent runs: analyze only new materials and append a new stage

Technical Approach

This project uses scripted analysis + AI semantic interpretation:

Scripts handle
- Scanning and filtering
- Time window decisions
- Stage detection and structured stats
- Output generation
AI handles
- Interpretation of structured results
- Stage signals and change hints
- Semantic fill-ins for designated sections

AI does not decide file selection, stage boundaries, or time ranges, keeping the process reproducible and traceable.

Project Structure

.claude/skills/personal-knowledge-growth-analyzer/
├── SKILL.MD                 # Skill behavior & constraints
├── scripts/                 # Core execution
│   ├── run_strategy.py      # Run strategy (first / incremental)
│   ├── scan_materials.py    # Scan & format detection
│   ├── detect_stage.py      # Stage detection
│   ├── analyze_stage.py    # Stage analysis
│   ├── update_baseline.py  # Baseline update
│   └── render_output.py    # Output generation
├── utils/
│   └── file_reader.py       # File reading helpers
├── templates/               # Markdown templates
│   ├── stage-growth-standard.md
│   └── stage-growth-canvas.md
└── analysis-output/         # Outputs (generated at runtime)

Design principles: