memory

Organize long-term facts, focus events, and decision history in memory files.

22|3|Updated Apr 22, 2026
One-click install
npx skills add https://github.com/MadeAgents/TopoClaw --skill memory-madeagents
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: memory
Source: https://github.com/MadeAgents/TopoClaw/tree/main/TopoClaw/topoclaw/skills/memory
Command: npx skills add https://github.com/MadeAgents/TopoClaw --skill memory-madeagents

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill prevents your assistant from forgetting important context by organizing long-term facts, tracking what you’re focused on over time, and preserving a searchable history of decisions.

Core Features & Use Cases

  • Layered workspace memory: maintains MEMORY (durable long-term facts), EVENT_MEMORY (structured focus and event timeline), and HISTORY (append-only log for review/search).
  • Focus-centric tracking: supports explicit vs inferred goals, time evolution (add/deepen/transfer/fade/undo), and consolidation of repeated small focuses into a compact timeline.
  • On-demand search strategy: chooses between in-memory filtering and targeted CLI grep/findstr/python search depending on file size.
  • Safe update boundaries: the compress_session tool trims the session tail only, while token-based archival can save chunks to MEMORY.md and HISTORY.md without automatically filling EVENT_MEMORY.

Quick Start

Ask the assistant to set up or update structured focus tracking by creating or editing memory/EVENT_MEMORY.md using the template for the current topic thread and focus (e.g., “track my current project focus and update the focus evolution log”).

Frequently Asked Questions about memory

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

FAQPage Schema
How do I keep assistant context and facts aligned across sessions?▼

Cross-session memory retention works by loading MEMORY.md and EVENT_MEMORY.md when non-empty, using an append-only HISTORY.md format with timestamped entries, and an internal save_memory pipeline for archival.

How do I track focus events and topic pivots over time?▼

Focus event tracking handles explicit and inferred goals with time evolution actions like add, deepen, transfer, fade, and undo, consolidating repeated small focuses into a compact timeline within EVENT_MEMORY.md.

Can I grep search my assistant's decision history?▼

Yes, you can grep search decision history by choosing between in-memory filtering and targeted CLI grep or findstr search depending on file size, querying the timestamped append-only entries in HISTORY.md.

What is the best way to archive long-term facts without losing session context?▼

The best way to archive long-term facts is using the compress_session tool to trim only the session tail, while token-based archival saves chunks to MEMORY.md and HISTORY.md without automatically filling EVENT_MEMORY.

Does workspace isolation support multiple concurrent project contexts?▼

Workspace isolation supports multiple concurrent project contexts by maintaining separate structured focus tracking and durable storage layers, preventing context bleed between different user preferences and topic threads.