Skip to content

Latest commit

 

History

History
52 lines (34 loc) · 1.46 KB

split-codex.md

File metadata and controls

52 lines (34 loc) · 1.46 KB
Raw

AI Codex Splitting Protocol

Invocation: @split-codex.md

Purpose

This protocol is used to intelligently split the AI Codex when it becomes too long, grouping related Errors (E) and Learnings (L) into separate codex pages.

Process

  1. Analyze the current @codex.md file
  2. Identify common themes or categories among entries
  3. Group related E and L entries
  4. Create new codex pages for each group
  5. Update the main @codex.md with links to new pages
  6. Report the new structure

Grouping Criteria

  • Project area (e.g., frontend, backend, database)
  • Technology or framework (e.g., React, Node.js, Prisma)
  • Concept (e.g., authentication, state management, testing)
  • File type or location (e.g., configuration files, utility functions)

File Structure

  1. Main codex remains at: .ai/codex/codex.md
  2. New pages: .ai/codex/[category]-codex.md

Update Process

  1. Move relevant entries to new category-specific files
  2. In the main codex (.ai/codex/codex.md), replace moved entries with links to new files
  3. Ensure cross-references are updated across all files

Naming Convention

Use kebab-case for new codex file names, e.g.:

  • frontend-codex.md
  • database-operations-codex.md
  • authentication-codex.md

Reporting

After splitting, provide a summary of:

  1. Number of new codex pages created
  2. Categories identified
  3. Entry distribution across new pages

IMPORTANT: Maintain the original entry IDs (E001, L001, etc.) in the new files to preserve existing references.