Claude Code Plugins

Community-maintained marketplace

Feedback
1
0

Use for all .md file creation/editing. README ≤200 lines (self-contained overview), topic files ≤400 lines. Plan ahead - if content will exceed limits, create directory structure from start.

Install Skill

1Download skill
2Enable skills in Claude

Open claude.ai/settings/capabilities and find the "Skills" section

3Upload to Claude

Click "Upload skill" and select the downloaded ZIP file

Note: Please verify skill by going through its instructions before using it.

SKILL.md

name progressive-disclosure
description Use for all .md file creation/editing. README ≤200 lines (self-contained overview), topic files ≤400 lines. Plan ahead - if content will exceed limits, create directory structure from start.

Progressive Disclosure for Markdown Documentation

Core Principle

README is self-contained - Contains all essential information, with optional links to details.

Not an index. Not a table of contents. A complete overview where you can stop reading and understand everything at that level.

The Rules

1. File Limits

  • README.md: ≤200 lines - Complete overview
  • {topic}.md: ≤400 lines - Detailed content on ONE topic
  • Plan ahead: If content will exceed limits, create directory structure from start

2. Anti-Spam Rule

Before creating new file, ask:

  1. Does this belong in existing file? → Update it
  2. Is this results of work already started? → Append to that file
  3. Only create new file for distinct new topic

3. README Structure Rule

Use topic sections with key findings + links (NOT generic "Subfiles:" lists)

✅ Do this:

## Clusivity

Inclusive/exclusive "we" distinctions affect 200+ languages. Austronesian languages
almost universally exhibit clusivity (kita/kami pattern). Validated: 100% accuracy.

[Read detailed analysis →](clusivity.md)

❌ Not this:

## Subfiles
- clusivity.md
- obviation.md

4. Link Format Rule

Use [text](file.md) - NOT aliases (aliases auto-import, breaking token efficiency)

Decision Tree: Creating Files

New content to add
    ↓
Can this go in existing file?
    ↓ YES → UPDATE that file
    ↓ NO
    ↓
Can this go in README?
    ↓ YES (README < 150 lines) → ADD to README
    ↓ NO
    ↓
Estimate content size
    ↓
Will this exceed 400 lines?
    ↓ YES → CREATE directory structure:
    |       mkdir {topic}/
    |       Create {topic}/README.md (overview)
    |       Create {topic}/{subtopic}.md files
    ↓ NO
    ↓
CREATE {topic}.md file

Key: Estimate BEFORE creating. Never create file knowing it will exceed 400 lines.

Templates

README.md Template

# {Topic Name}

{2-3 sentence overview}

## {Subtopic 1}

{KEY FINDINGS - 2-4 sentences with actual insights, not just "see file"}

**Status:** {done/in progress}

[Read detailed analysis →]({subtopic-1}.md)

## {Subtopic 2}

{KEY FINDINGS}

**Status:** {status}

[Read detailed analysis →]({subtopic-2}.md)

Rules:

  • Essential findings IN readme (don't force clicking)
  • Links for DETAILS, not basic understanding
  • Under 200 lines total

Topic File Template

# {Topic}

> **Parent Context:** {1 sentence from parent README}

{Introduction}

## {Section 1}

{Content}

## Summary

{Key takeaways}

Experiment Template (All in ONE file)

# Experiment 001: {Title}

## Hypothesis
{What you think}

## Method
{How you test}

## Results
{What happened} ← APPEND after running

## Learnings
{What you learned} ← APPEND after analyzing

## Next Steps
{What's next} ← APPEND after reflecting

Do NOT create: hypothesis.md, method.md, results.md separately!

Common Patterns

Pattern 1: Simple Feature Research

features/person-systems/
├── README.md (overview + findings for each topic)
├── clusivity.md
└── obviation.md

README.md:

# Person Systems

Person systems (clusivity, obviation, number) affect 200+ languages.

## Clusivity

Inclusive/exclusive "we" distinctions. Universal in Austronesian languages.
Validated: 100% accuracy on 12 verses.

[Read detailed analysis →](clusivity.md)

## Obviation

Fourth person systems. Common in Algic languages.
Framework complete, needs validation.

[Read detailed analysis →](obviation.md)

Pattern 2: Nested (when clusivity.md would exceed 400 lines)

features/person-systems/
├── README.md
├── clusivity/              # Started as directory (content >400 lines)
│   ├── README.md           # Clusivity overview
│   ├── validation.md       # All validation experiments
│   └── patterns.md         # Language patterns
└── obviation.md            # Still under 400 lines

clusivity/validation.md - NOT individual verse files!

# Clusivity Validation

## Divine Speaker Verses

### Genesis 1:26 "Let us make mankind"
Prediction: Exclusive
Results: Indonesian kami ✓, Tagalog kami ✓

### Genesis 3:22 "Man has become like us"
[same format]

## Apostolic Authority Verses
[grouped verses]

## Summary
84 predictions, 84 correct (100%)

Pattern 3: Experiments

experiments/aspect/
├── README.md
├── experiment-001.md  # Hypothesis → Method → Results → Learnings (all in one)
└── experiment-002.md

experiment-001.md:

# Experiment 001: Basic Aspect

## Hypothesis
Aspect predictable from morphology

## Method
20 verbs from John 1

## Results
70% accuracy. Failed on perfect aspect.

## Learnings
- Aorist: 100% accurate
- Perfect needs semantic context
- Changes for 002: add context window

Pattern 4: Bible Study Tools

bible-study-tools/cultural-context/
├── README.md (≤200 lines)
│   ## Purpose
│   ## Experiments
│   {Summary with findings}
│   [Read experiments →](experiments.md)
│   ## Schema
└── experiments.md (≤400 lines)
    ## Experiment 001
    {Hypothesis → Method → Results → Learnings}

Only create experiments/ directory if experiments.md exceeds 400 lines.

Anti-Patterns

❌ File Spam

experiments/
├── experiment-001-hypothesis.md
├── experiment-001-results.md
└── experiment-001-learnings.md

✅ Correct: experiment-001.md (all sections in one file)

❌ Random Verse Files

clusivity/
├── GEN-001-026.md
├── MAT-006-009.md

✅ Correct: validation.md (organized verse sections)

❌ Generic Subfile Lists

## Subfiles
- file1.md
- file2.md

✅ Correct: Topic sections with findings + links

❌ QUICK-REFERENCE.md

├── README.md (long)
└── QUICK-REFERENCE.md (short)

✅ Correct: README IS the quick reference (≤200 lines)

❌ Creating Oversized Files

Agent: "I'll create validation.md with 622 lines, then note it should be a directory"

✅ Correct: Estimate first. If >400 lines, create validation/ directory from start

Workflow

Creating New Documentation

  1. Start simple: Create directory with minimal README.md
  2. Grow README: Add content directly as you learn
  3. Before adding more: Will README exceed 200 lines?
    • YES → Extract to topic file, summarize in README with link
    • NO → Keep adding to README
  4. Before creating topic file: Will it exceed 400 lines?
    • YES → Create directory structure instead
    • NO → Create topic.md file

Estimating File Size

Before creating topic file:

  • Count major sections planned
  • Estimate lines per section
  • Add 20% buffer for growth

Examples:

  • 3 sections × 100 lines = 300 lines → Create file ✓
  • 5 sections × 100 lines = 500 lines → Create directory ✗
  • Unsure? → Create file, watch line count, convert if approaching 400

Converting File to Directory

If topic.md approaches 400 lines:

  1. Create topic/ directory
  2. Create topic/README.md with overview
  3. Split content into topic/{subtopic}.md files
  4. Update parent README link

Quick Reference

Creating markdown? Ask:

  1. Can this go in existing file? → UPDATE it
  2. Is README under 150 lines? → ADD to README
  3. Is related file under 300 lines? → ADD to that file
  4. Creating new topic file - will it exceed 400 lines? → CREATE DIRECTORY instead
  5. Otherwise → CREATE FILE

README must have:

  • ✅ Essential findings (not just links)
  • ✅ Topic sections with key info + links
  • ✅ Under 200 lines
  • ❌ NOT just a file list
  • ❌ NO "Subfiles:" section

File creation:

  • ✅ Plan ahead - estimate size first
  • ✅ One distinct topic = one file
  • ✅ Append sections as work progresses
  • ✅ Update existing before creating new
  • ❌ NO oversized files (>400 lines)
  • ❌ NO separate files for parts of same work
  • ❌ NO file spam