Agent Memory

A lightweight CLI for managing daily memory files as an AI agent. It handles the boring stuff like checking for gaps in your daily logs, searching across files with context, and creating today's entry if you forgot. The digest command is the interesting bit: it uses Gemini to summarize the last week of memories into something you can actually review. Sensible defaults for memory directory location, and the README shows good integration hygiene with heartbeat checks. This is basically housekeeping automation for agents that keep markdown journals. If you're already doing daily logs manually, this removes the friction.

skills.sh installs: 341
last updated: Jun 3, 2026

View on GitHub

Install

$ npx -y skills add molty-assistant/agent-memory-skill --skill agent-memory --agent claude-code

Installs into .claude/skills of the current project.

Files

Agent Memory (memctl)

Memory management CLI for AI agents. Organize, search, and maintain your memory files.

Install

npm install -g agent-memory

Or clone and build:

git clone https://github.com/molty-assistant/agent-memory.git
cd agent-memory && npm install && npm run build

Commands

List memory files

memctl list              # List all
memctl ls --recent 5     # Show recent

Search across files

memctl search "query"              # Find mentions
memctl s "project" --context 3     # With context lines

Summary stats

memctl summary           # Last 7 days stats
memctl sum --days 30     # Last 30 days

Check for gaps

memctl gaps              # Missing daily entries (30 days)
memctl gaps --days 7     # Check last week

Create today's file

memctl touch             # Creates YYYY-MM-DD.md if missing

AI-powered digest (requires Gemini API key)

export GEMINI_API_KEY=your_key
memctl digest            # AI summary of last 7 days
memctl ai --days 3       # Last 3 days
memctl digest -o out.md  # Save to file