authenticwalk avatar

bible-study-tool-creator

Interactive skill for creating new bible study tools that generate AI-readable commentary data. Guid

by authenticwalk|Open Source

Bible Study Tool Creator Skill

Overview

This skill helps you create new bible study tools for the myBibleToolbox project. It provides a complete workflow for:

  1. Checking for duplicate tools
  2. Gathering tool requirements from users
  3. Verifying practical value through concrete examples
  4. Generating complete file structures
  5. Setting up self-learning loops for quality improvement

What Gets Created

When you use this skill to create a new tool, it generates:

/bible-study-tools/{tool-name}/
├── README.md                      # Tool documentation with goals and examples
├── LEARNING.md                    # Experiment log for tracking improvements
├── {tool-name}-template.md        # Agent prompt template for data generation
└── tests/                         # Self-learning loop test framework
    └── README.md                  # Test documentation

Files in This Skill

SKILL.md

The main skill definition that Claude uses. Contains step-by-step workflow for creating new tools.

init-tool.py

Python script that generates all files from a YAML definition. Takes user input and populates templates.

templates/

  • user-input.template.md: Interactive form for gathering tool requirements
  • README.template.md: Template for tool README
  • LEARNING.template.md: Template for experiment tracking
  • tool-template.template.md: Template for agent prompts
  • tests-README.template.md: Template for test framework documentation
  • analysis.template.md: Template for test output analysis

How to Use

Option 1: Via Claude Skill

Invoke the skill directly:

Hey Claude, I want to create a new bible study tool using bible-study-tool-creator

Claude will:

  1. Check for existing similar tools
  2. Present you with a template to fill out
  3. Review your examples for practical value
  4. Generate all files automatically
  5. Give you next steps for the self-learning loop

Option 2: Manual Script Execution

If you already have a tool definition:

python3 .claude/skills/bible-study-tool-creator/init-tool.py /path/to/tool-definition.yaml

Tool Definition Format

The YAML file should follow this structure:

tool_name: "Greek Word Analysis"
tool_name_kebab: "greek-word-analysis"
task_name: "greek-words"
description: "One sentence description"
test_verse: "MAT 5:3"

goals_formatted: |
  1. Goal one
  2. Goal two
  3. Goal three

examples_formatted: |
  ### Example 1
  **Context**: ...
  **Insight**: ...
  **Value**: ...

  ### Example 2
  ...

related_tools: "tool-name-1, tool-name-2"

data_structure: |
  ```yaml
  verse:
    reference: "MAT-5-3"
    your_data:
      field1: "..."
      field2: "..."

yaml_structure_inline: | your_data: field1: "..." field2: "..."


## Requirements Philosophy

### The 5 Examples Rule

Every tool MUST have 5 concrete examples showing:
- **Context**: What verse/concept is being analyzed
- **Insight**: What the tool reveals
- **Value**: How this helps translators/pastors/students

Why 5? Because if you can't find 5 compelling examples, the tool might not be valuable enough to build.

### Practical Value Focus

Goals must be grounded in helping real people (Bible translators, pastors, students) use AI more effectively. Avoid:
- Data that's easy for LLMs to remember (basic facts)
- Data for data's sake
- Theoretical or vague objectives

Focus on:
- Translation challenges
- Cultural context
- Interpretive nuances
- Meaningful connections

## Self-Learning Loop

Each tool is designed for continuous improvement:

1. **Test**: Run agent with template on a verse
2. **Analyze**: Critically evaluate output quality
3. **Learn**: Document findings in LEARNING.md
4. **Refine**: Improve the template
5. **Repeat**: Test again until excellent

This ensures the generated data is:
- Accurate
- Useful
- Complete
- Unique (not just regurgitating LLM training data)

## Examples of Good vs Bad Tools

### ✅ Good Tool: Cultural Translation Adaptations

**Why it's good**:
- Addresses real translation challenge (culturally-specific words)
- 5 strong examples (snow → egret feathers, sheep → caribou, etc.)
- Clear value for translators
- Data not easily remembered by LLMs

### ❌ Bad Tool: List All Books of the Bible

**Why it's bad**:
- LLMs already know this
- No practical translation/interpretation value
- Doesn't need a "book's worth of context"
- Not grounded in the project's mission

## Contributing

If you improve this skill:
1. Update templates in `templates/` directory
2. Update `SKILL.md` with new workflow steps
3. Update `init-tool.py` if changing data structure
4. Document changes in this README

## Related Documentation

- [Project README](../../../README.md) - Overall project goals and task ideas
- [CLAUDE.md](../../../CLAUDE.md) - Project structure and conventions
- [quote-bible skill](../quote-bible/SKILL.md) - Example of well-structured skill

## License

MIT License - same as the parent project
Download Skill