memory

Name: memory
Availability: InStock
Author: Ethan

Retrieve long-term facts and event logs via grep-based searches.

65|9|Updated Jan 29, 2026

Search Tags:#context #recall #memory #persistence #history #grep #long-term

Ready to run

npx skills add https://github.com/EthanAlgoX/MarketBot --skill memory-ethanalgox

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: memory
Source: https://github.com/EthanAlgoX/MarketBot/tree/main/marketbot/skills/memory
Command: npx skills add https://github.com/EthanAlgoX/MarketBot --skill memory-ethanalgox

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Provides a lightweight, reliable two-layer memory so the assistant can keep important long-term facts in context while enabling fast searches over an append-only event log.

Core Features & Use Cases

Two-layer storage: memory/MEMORY.md holds persistent facts and is always loaded into context, while memory/HISTORY.md is an append-only event log not loaded by default.
Grep-based recall: Use the exec tool to run grep on HISTORY.md for quick retrieval of past events without bloating the active context.
Write and consolidate: Important facts should be written immediately with edit_file or write_file, and older conversations are auto-summarized and consolidated into the history.
Use Case: Preserve user preferences and project facts across sessions and quickly locate past meeting notes, deadlines, or decisions.

Quick Start

Search memory/HISTORY.md for events mentioning deadline and return the most relevant entries.

Frequently Asked Questions about memory

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

FAQPage Schema

How do I add persistent memory to an AI assistant for long-term context recall?▼

Persistent memory for long-term context recall uses a two-file structure where persistent facts are stored in MEMORY.md and loaded into context, while an append-only HISTORY.md log captures events for grep-based retrieval.

What is the best way to search past conversation history without bloating the active context window?▼

The best way to search past conversation history without context bloat is using an append-only event log like HISTORY.md, which stays unloaded by default and supports grep-based retrieval for fast, targeted recall of specific events.

How does append-only event logging work for tracking project decisions and deadlines?▼

Append-only event logging for tracking project decisions and deadlines works by continuously writing new entries to a HISTORY.md file, allowing you to grep past records without modifying existing data or loading the entire log into context.

Can I use grep to retrieve specific user preferences from a persistent memory log?▼

Yes, you can use grep to retrieve specific user preferences by executing the search tool directly on the HISTORY.md file, quickly locating past preference records without loading the entire event log into the active context.

When should I consolidate conversation history into persistent memory files?▼

You should consolidate conversation history into persistent memory files immediately when important facts arise, writing them to MEMORY.md or HISTORY.md to ensure user preferences and project facts are preserved across sessions.

Does this two-layer memory approach require loading the entire event history into the context window?▼

No, the two-layer memory approach keeps the event history file unloaded by default to prevent context bloat, relying instead on grep commands to fetch only the relevant historical entries when recall is needed.