memory

Store stable facts in memory and isolate transient data for retrieval.

Updated Apr 9, 2026

Search Tags:#context #search #memory #persistence #archive #grep #history.md

Ready to run

npx skills add https://github.com/HuaGCS/Hahobot --skill memory-huagcs

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: memory
Source: https://github.com/HuaGCS/Hahobot/tree/main/hahobot/skills/memory
Command: npx skills add https://github.com/HuaGCS/Hahobot --skill memory-huagcs

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Two-layer memory system that preserves long-term facts and event history while enabling fast, grep-based search over archived observations.

Core Features & Use Cases

Long-term facts are stored in memory/MEMORY.md and automatically loaded into the active context to maintain project context and preferences.
Append-only event history is kept in memory/HISTORY.md to enable chronological reconstruction and auditing, with grep-style search as needed.
Structured archive data under memory/archive/ (index.jsonl and chunks/*.json) supports history search, timeline, and expansion across sessions.

Quick Start

Load the memory files into your context and begin recording memory entries to test retrieval and search.

Frequently Asked Questions about memory

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

FAQPage Schema

How do I persist long-term context across sessions for a knowledge-heavy workflow?▼

To persist long-term context across sessions, store stable project facts in a memory file that automatically loads into the active context. This maintains project preferences and ensures continuity throughout your knowledge-heavy workflows.

What is the best way to search an append-only event history in markdown?▼

The best way to search an append-only event history in markdown is using grep-style search over a dedicated history file. This enables chronological reconstruction and auditing of events directly from your stored text data.

How does a structured memory archive support timeline reconstruction?▼

A structured memory archive supports timeline reconstruction by indexing observations into JSONL files and JSON chunks. This allows you to search historical data, expand timelines, and access archived context across multiple sessions.

Can I isolate transient data from stable facts in a two-layer memory system?▼

Yes, a two-layer memory system isolates transient data from stable facts by storing long-term preferences separately from append-only event history. This prevents temporary context from overwriting persistent project knowledge.

Does the memory system require external dependencies for grep-based search?▼

No, the memory system does not require external dependencies for grep-based search. It relies entirely on native text files and structured JSON archives to perform fast, grep-style retrieval over stored observations.

What file formats are used to store structured memory archives?▼

Structured memory archives are stored using Markdown files for facts and history, alongside JSONL index files and JSON chunks for structured data. These formats support fast grep-style search and cross-session retrieval.