Can I manually edit the memory file?

It is highly recommended to use the provided Python scripts (add_memory_entry.py, edit_memory_entry.py) to maintain file integrity, sequential IDs, and consistent formatting.

What is the main purpose of Memory Curator?

It creates a durable, machine-readable log of architectural decisions and lessons learned to prevent knowledge loss and provide context for future development tasks.

When should I create a new memory entry?

You should create an entry if you've made an important architectural decision, solved a non-obvious problem worth remembering, or gained an insight that would be useful for someone else working on the project later.

How does it store memory entries?

Entries are stored in a .meridian/memory.jsonl file as single-line JSON objects, which ensures the file remains easily parsable by both developers and AI tools.

Memory Curator

Name: Memory Curator
Author: markmdev

bymarkmdev

•

118

Learning & Documentation

Captures and manages durable architectural decisions and engineering insights within a structured memory log.

About

Memory Curator is a specialized skill for Claude Code designed to build a persistent audit trail of engineering knowledge. By leveraging a structured memory.jsonl format, it allows developers to document critical architectural patterns, lessons learned from production incidents, and the rationale behind complex technical decisions. This ensures that vital project context remains accessible and actionable, preventing knowledge loss during context window compaction or long-term development cycles. The skill enforces high-quality documentation standards through automated scripts that handle sequential ID generation, timestamping, and metadata tagging.

Key Features

Standardized engineering audit trail via memory.jsonl
Structured metadata tagging for improved searchability
Safe update and deletion operations via CLI helper scripts
118 GitHub stars
Automated entry creation with sequential IDs and UTC timestamps
Cross-referencing between memory entries, task IDs, and source files

Use Cases

Recording post-mortem insights from production incidents to avoid recurring architectural mistakes
Documenting why a specific library or framework was chosen over alternatives to prevent future regression
Maintaining long-term architectural context for new team members or AI agents working on the codebase

About

Key Features

Standardized engineering audit trail via memory.jsonl
Structured metadata tagging for improved searchability
Safe update and deletion operations via CLI helper scripts
118 GitHub stars
Automated entry creation with sequential IDs and UTC timestamps
Cross-referencing between memory entries, task IDs, and source files

Use Cases

Recording post-mortem insights from production incidents to avoid recurring architectural mistakes
Documenting why a specific library or framework was chosen over alternatives to prevent future regression
Maintaining long-term architectural context for new team members or AI agents working on the codebase