← ClaudeAtlas

agentmind-memorylisted

Persistent cross-session memory system for AI agents. Four-layer architecture (Working/Episodic/Semantic/Procedural) stored as local Markdown files — human-readable, human-editable, zero dependencies. Your agent remembers user preferences, project context, lessons learned, and daily events across conversations. Pairs with agentmind-metacognition for a complete "brain OS."
Ahmed-Mazh2r/AgentMind · ★ 1 · AI & Automation · score 77
Install: claude install-skill Ahmed-Mazh2r/AgentMind
# 🧠 AgentMind Memory v1.1 > **Your agent forgets everything when the conversation ends. This fixes that.** > **All memory stored as Markdown files. No database. No API. No vendor lock-in.** --- ## What This Solves Every AI agent has the same problem: start a new conversation, and it's a stranger again. It doesn't know your name, your projects, your preferences, or the mistakes it made yesterday. AgentMind Memory gives your agent a local, persistent brain: - **Remembers** your preferences and project context across sessions - **Learns** from mistakes and never repeats the same error - **Tracks** what happened when, so nothing falls through the cracks - **Stays yours** — plain Markdown files on your machine, fully editable --- ## Architecture Four memory layers, inspired by cognitive science: ``` ┌─────────────────────────────────────────────────────────┐ │ AgentMind Memory │ ├──────────────┬──────────────┬──────────────┬────────────┤ │ WORKING │ EPISODIC │ SEMANTIC │ PROCEDURAL │ │ │ │ │ │ │ Current task │ Daily event │ Long-term │ How to do │ │ state & │ logs with │ facts & │ things & │ │ progress │ timestamps │ preferences │ lessons │ │ │ │ │ learned │ ├──────────────┼──────────────┼──────────────┼────────────┤ │ WORKING.md │ memory/ │ MEMORY.md │ PROCS.md │ │ │ YYYY-MM