Agent Skills: CLAUDE.md Updater

Updates CLAUDE.md based on recent conversation history. This skill should be used when the user asks to update their CLAUDE.md, sync learnings from conversations, or review what was discussed recently.

UncategorizedID: yurifrl/cly/ag:claudemd-update

Install this agent skill to your local

pnpm dlx add-skill https://github.com/yurifrl/cly/tree/HEAD/.agents/skills/ag%3Aclaudemd-update

Skill Files

Browse the full folder contents for ag:claudemd-update.

Download Skill

Loading file tree…

.agents/skills/ag:claudemd-update/SKILL.md

Skill Metadata

Name
ag:claudemd-update
Description
Updates CLAUDE.md based on recent conversation history. This skill should be used when the user asks to update their CLAUDE.md, sync learnings from conversations, or review what was discussed recently.

CLAUDE.md Updater

Extract Conversations

scripts/extract_conversations.py --project "$(pwd)" --days 1

Options: --days N, --limit N, --json

Workflow

  1. Extract conversations - Execute the script immediately:

    scripts/extract_conversations.py --project "$(pwd)" --days 1 --limit 15
    
  2. Extract final state - Analyze conversations for the end result:

    • Final decisions and patterns that emerged
    • Current conventions and rules (not how they were discussed)
    • Workflow steps as they should be done now
    • Configuration patterns that were settled on

    CRITICAL: Focus on the final product, not the discussion history. Skip:

    • Intermediate steps or abandoned approaches
    • Debugging sessions (unless general pattern)
    • Version-specific details
    • One-time fixes
  3. Present clean documentation - Show numbered list of rules/patterns to add:

    Patterns to document:
    1. [Final pattern/rule as it should be done]
    2. [Configuration pattern with current values]
    3. [Workflow step in present tense]
    ...
    
    Which should I document? (numbers, ranges, or "all")
    
  4. Apply updates - After selection, write clean documentation focused on current state