memory

Store long-term facts in MEMORY.md and searchable HISTORY.md event logs.

37|1|Updated Mar 17, 2026
One-click install
npx skills add https://github.com/saolalab/clawforce --skill memory-saolalab
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: memory
Source: https://github.com/saolalab/clawforce/tree/main/clawbot/skills/memory
Command: npx skills add https://github.com/saolalab/clawforce --skill memory-saolalab

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Agents often lose context over time, making it hard to remember long‑term facts or retrieve past events without manual searching.

Core Features & Use Cases

  • Two‑layer memory: Persistent MEMORY.md stores always‑loaded facts, while HISTORY.md records an append‑only event log.
  • Grep‑based search: Quickly locate relevant past events using standard grep commands through the exec tool.
  • Auto‑consolidation: Conversations are summarized into HISTORY.md and distilled into MEMORY.md automatically, requiring no manual maintenance.
  • Easy updates: Add new facts instantly with edit_file or write_file to keep the long‑term memory current.

Quick Start

Ask the memory skill to find all mentions of “deadline” in recent events.

Frequently Asked Questions about memory

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

FAQPage Schema
How do I add long-term memory to an autonomous agent workflow?

Long-term memory for autonomous agents is managed through a two-layer architecture using MEMORY.md for always-loaded facts and HISTORY.md for append-only event logs, providing persistent context without manual maintenance.

How can I search past event logs for an AI agent?

Past event logs are searchable using standard grep commands executed through the exec tool, allowing you to quickly locate specific mentions or events within the HISTORY.md append-only log file.

Does agent memory auto-consolidation require manual summarization?

Auto-consolidation requires no manual maintenance, as conversations are automatically summarized into HISTORY.md and distilled into MEMORY.md, ensuring long-term facts and event logs stay current automatically.

What is the best way to store persistent context for AI agents?

Storing persistent context is best handled by a two-layer memory architecture using MEMORY.md for always-loaded facts and HISTORY.md for searchable logs, which automatically consolidates conversations to retain context.

Can I update long-term facts in MEMORY.md manually?

Long-term facts in MEMORY.md can be updated instantly using edit_file or write_file commands; while auto-consolidation handles automatic distillation, direct file edits keep the memory current.