memory-manager

Store and retrieve cross-session decisions, error patterns, and project knowledge in YAML files.

Updated Aug 27, 2026

Search Tags:#knowledge management #memory #decision logging #project continuity #marvin #error patterns

Ready to run

npx skills add https://github.com/brunoldqueiroz/marvin --skill memory-manager-brunoldqueiroz

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: memory-manager
Source: https://github.com/brunoldqueiroz/marvin/tree/main/.claude/skills/memory-manager
Command: npx skills add https://github.com/brunoldqueiroz/marvin --skill memory-manager-brunoldqueiroz

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill provides a robust system for Marvin to remember past decisions, error patterns, and project knowledge across different sessions, preventing repeated mistakes and accelerating development.

Core Features & Use Cases

Persistent Memory: Stores and retrieves information across multiple Marvin sessions, building a project-specific knowledge base.
Decision & Error Tracking: Logs architectural decisions, rejected alternatives, and recurring error patterns with confidence scores.
Knowledge Map: Maintains an up-to-date understanding of project structure and conventions.
Use Case: If Marvin previously decided on a specific module structure for a Python project, this skill ensures that decision is recalled and applied in future tasks related to that project, avoiding re-litigation of the same architectural choices.

Quick Start

Use the memory manager skill to store the decision about using Qdrant for the knowledge base.

Frequently Asked Questions about memory-manager

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I persist decisions and error patterns across Marvin development sessions?▼

Cross-session memory for Marvin works by storing architectural choices, error patterns, and project conventions in structured files with YAML frontmatter. This enables proactive recall of past knowledge, ensuring continuity and learning across development sessions.

What is the best way to log architectural decisions and rejected alternatives in a project?▼

Logging architectural decisions and rejected alternatives requires a structured, file-based system with typed subdirectories and YAML frontmatter. By recording choices with confidence scores, this tiered memory architecture ensures past decisions are recalled and applied in future tasks.

Does Marvin memory support proactive recall of project-specific conventions and error patterns?▼

Yes, proactive recall of project-specific conventions and error patterns is supported through a tiered memory architecture. It uses typed subdirectories and YAML frontmatter for efficient querying, maintaining an up-to-date understanding of project structure and conventions.

How do I maintain a knowledge map of project structure and conventions for Marvin?▼

Maintaining a knowledge map of project structure and conventions involves storing and retrieving project knowledge in a structured, file-based system. This ensures Marvin maintains an up-to-date understanding of your project, facilitating continuity and learning across development sessions.

Can I use file-based YAML frontmatter to store cross-session project knowledge for AI assistants?▼

Yes, using file-based YAML frontmatter to store cross-session project knowledge is supported. It organizes decisions, error patterns, and project conventions in typed subdirectories, enabling efficient querying and proactive recall across different development sessions.

Why does Marvin forget architectural decisions between different development sessions?▼

Marvin forgets architectural decisions between sessions because it lacks a persistent cross-session memory system. Without a structured, file-based memory architecture with typed subdirectories, past decisions, error patterns, and project knowledge are not retained or recalled.