workspace-memory

Maintains a persistent topic-indexed memory graph of workspace decisions and events.

Updated Aug 6, 2026
One-click install
npx skills add https://github.com/ferrarifankid04/ai-skills-public --skill workspace-memory-ferrarifankid04
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: workspace-memory
Source: https://github.com/ferrarifankid04/ai-skills-public/tree/main/claude-code/skills/workspace-memory
Command: npx skills add https://github.com/ferrarifankid04/ai-skills-public --skill workspace-memory-ferrarifankid04

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI assistants forget context between sessions, forcing users to re-explain past decisions, file exports, and project history. This Skill maintains a persistent, topic-organized memory graph inside the workspace so prior work can be recalled instantly. ## Core Features & Use Cases - Automatic Memory Capture: Appends one-line dated entries with source links to topic files whenever a file is created, a decision is made, or the user says "remember". - Navigable Graph Structure: Uses a .memory/index.md table of contents linking to individual topic pages under .memory/topics/ for gradual disclosure. - Silent Operation: Updates memory without narrating the process, only confirming saves with a single plain-English sentence. - Use Case: After exporting stargazer data last week, ask "what did we decide about the stargazer export?" and the assistant reads the relevant topic page and answers with a link to the original discussion. ## Quick Start Ask the assistant to remember a decision you just made, then in a later session ask it to recall what was decided about that topic.

Frequently Asked Questions about workspace-memory

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

FAQPage Schema
How do I give an AI assistant persistent memory across sessions?

Use a file-based memory graph stored in the workspace itself. This Skill writes dated one-line entries to topic pages under .memory/topics/ and maintains an index.md table of contents, so context persists between sessions without external databases.

How does the workspace memory graph organize information?

It uses a two-level structure: an index.md file with one line per topic linking to individual topic pages in .memory/topics/. Each topic page holds chronological one-line entries with a date, brief description, and a link to the original discussion.

What events get saved to workspace memory automatically?

Memory is updated when a file is created or significantly modified, a decision or direction is chosen, a workflow produces durable output, or the user explicitly says remember or note. Conversational chatter, general knowledge, and credentials are never captured.

Can I use workspace memory with sensitive credentials or secrets?

No, credentials should not be stored. The Skill explicitly excludes credentials and actively edited content from capture, recording only durable decisions, outputs, and events as short dated entries.

Why does the assistant not find my previous memory entries?

Retrieval always starts from .memory/index.md to locate the correct topic page before reading it. If the index is empty or missing, memory was never bootstrapped, so new entries must be created first.