jrnl-mcp: Instructions file for Claude Code

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

📋 Work Tracking - START HERE

IMPORTANT: Before starting any work, check these documents in order:

1. docs/roadmap.md - 3-phase improvement plan (Phase 1 → Phase 2 → Phase 3)

   • Overview of all planned improvements
   • Categorized by priority (high/medium/low)
   • Helps understand the big picture

2. docs/todo.md - Detailed Phase 1 tasks with work logs

   • Most important for day-to-day work
   • Detailed task breakdowns with checkboxes
   • Work logs with timestamps and results
   • Next person reads this to continue work

3. This file (CLAUDE.md) - Project overview and guidelines

   • Architecture and implementation requirements
   • Development guidelines and best practices

---

Git Workflow - MUST READ

CRITICAL: The main branch is protected. Direct pushes are blocked.

Before ANY Git Operation

git branch  # Always check current branch first

Required Workflow

develop (work here) → PR → main (protected)

1. Always work on develop branch

   git checkout develop
   

2. Commit and push to develop

   git add <files>
   git commit -m "message"
   git push origin develop
   

3. Create PR to merge into main

   gh pr create --base main --head develop
   

4. Never commit directly to main - It will be rejected by branch protection rules.

If You Accidentally Commit to Main

# Cherry-pick to develop
git checkout develop
git cherry-pick <commit-hash>
git push origin develop

# Reset main to remote
git checkout main
git reset --hard origin/main

---

Project Overview

This is a jrnl MCP (Model Context Protocol) server that provides a read-only API for AI assistants to access and analyze journal entries from the jrnl command-line tool.

Current State: Phase 1 improvements in progress (see docs/todo.md for details).