Skip to content

Conversation

shinpr
Copy link
Owner

@shinpr shinpr commented Sep 10, 2025

Summary

  • Added comprehensive rule editing guide with 9 LLM optimization principles
  • Enhanced use-cases documentation with clearer explanations and practical examples
  • Improved overall documentation clarity to address user confusion about project philosophy

Changes

New Documentation

  • rule-editing-guide.md: Comprehensive guide for creating effective rules that maximize LLM execution accuracy
    • 9 core principles for rule optimization
    • Practical examples with before/after comparisons
    • Troubleshooting section for common issues
    • Mapping of principles to existing rule files

Enhanced Documentation

  • use-cases.md:

    • Added clearer command explanations with "what happens" descriptions
    • Included phase-based implementation examples
    • Added recovery instructions for interrupted tasks
    • Improved terminology consistency (AI → LLM)
    • Enhanced descriptions of Agentic Coding and meta-cognition concepts
  • sync-rules.md:

    • Added judgment criteria to prevent unnecessary modifications
    • Clarified when updates are actually needed vs. when to skip
  • README.ja.md:

    • Added link to rule editing guide
    • Updated directory structure documentation

shinpr and others added 2 commits September 10, 2025 13:54
Remove /onboard command entirely based on analysis showing:
- Conflicts with current architecture philosophy (minimal context preparation)
- Low actual usage despite documentation presence
- Better alternatives exist (rule-advisor for targeted rule selection)
- Simplifies command set and reduces maintenance overhead

Changes:
- Delete command files: .claude/commands*/onboard.md
- Update README workflow: remove /onboard step
- Update use-cases guides: remove /onboard documentation
- Clean troubleshooting references to /onboard

The sub-agent approach with targeted rule selection via rule-advisor
provides better alignment with the "single responsibility" and
"just-in-time context generation" design principles.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
- Enhanced use-cases.md with clearer command explanations and practical examples
- Added comprehensive rule-editing-guide.md with 9 LLM optimization principles
- Updated sync-rules command to prevent unnecessary modifications
- Added links to rule editing guide in README.ja.md
- Clarified Agentic Coding concepts and sub-agent importance throughout docs
- Added troubleshooting sections and phase-based implementation examples
- Improved terminology consistency (AI → LLM) for technical accuracy

These changes address user confusion about project philosophy and provide
better guidance for customizing rules to maximize LLM execution accuracy.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
@shinpr shinpr self-assigned this Sep 10, 2025
shinpr and others added 3 commits September 10, 2025 21:01
- Translated rule-editing-guide.md to English with LLM optimization principles
- Updated use-cases.md with enhanced command explanations and practical examples
- Added decision criteria to sync-rules command to prevent unnecessary changes
- Updated README.md with links to rule editing guide in both languages
- Adjusted token size recommendations for English (250 lines ≈ 1,500 tokens)
- Improved terminology consistency across all English documentation
- Added troubleshooting sections and phase-based implementation examples

These translations ensure English-speaking users have access to the same
comprehensive documentation improvements as Japanese users.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
- Renamed section from "Important Decision Criteria" to "Change Necessity Evaluation"
- Removed excessive emphasis that misrepresented priority hierarchy
- Optimized English version for LLM execution accuracy with IF-THEN-TERMINATE structure
- Maintained measurable criteria (100% synchronization, zero mismatches)
- Balanced emphasis levels to properly reflect metadata sync as primary responsibility

The change evaluation is now appropriately positioned as a conditional branch
rather than a critical component, ensuring correct prioritization of the
command's main purpose.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
- Reorganize both English and Japanese READMEs with clearer structure
- Add table of contents for easier navigation
- Move sub agents list to the end, show only key agents
- Emphasize real-world results with time comparisons
- Simplify workflow explanations for better understanding
- Add demo GIF as H3 section under "Why Sub Agents"
- Use more approachable tone in Japanese version
- Focus on practical benefits over technical details

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
@shinpr shinpr marked this pull request as ready for review September 10, 2025 13:01
@shinpr shinpr merged commit d7655a4 into main Sep 10, 2025
1 check passed
@shinpr shinpr deleted the improve-documentation-clarity branch September 10, 2025 13:02
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant