Gemini CLI Commands for Forgetful

Custom commands for using Forgetful with Gemini CLI.

Credits: These commands were contributed by @CharlieBytesX in Discussion #15.

Prerequisites

Ensure Forgetful MCP server is configured:

# STDIO transport (recommended for local use)
gemini mcp add forgetful uvx forgetful-ai

# Or HTTP transport (for remote/Docker)
gemini mcp add -t http forgetful http://localhost:8020/mcp

Installation

Copy all commands to your Gemini CLI commands directory:

cp docs/gemini-cli/commands/*.toml ~/.gemini/commands/

Or copy individual commands as needed.

Available Commands

Command	Description
/encode-repo	Bootstrap a repository into Forgetful's knowledge base using a 7-phase protocol
/forgetful-setup	Configure the Forgetful MCP server
/memory-explore	Deep exploration of the Forgetful knowledge graph
/memory-list	List recent memories from Forgetful
/memory-save	Save current context as an atomic memory with curation workflow
/memory-search	Search memories semantically

Usage Examples

Encode a Repository

Bootstrap a new project into Forgetful:

/encode-repo my-project

The command walks through 7 phases: discovery, foundation, architecture, patterns, features, decisions, and validation.

Save a Memory

After a conversation where you learned something important:

/memory-save

The command will analyze the conversation, check for existing related memories, and propose a new atomic memory with proper curation.

Search Memories

Find relevant context from your knowledge base:

/memory-search authentication patterns

List Recent Memories

See what's been added recently:

/memory-list 10

Explore the Knowledge Graph

Deep traversal of related memories, entities, and documents:

/memory-explore payment processing

Command Format

Gemini CLI commands are TOML files supporting:

• {{args}} - Argument injection from user input
• @{path} - File content injection
• !{command} - Shell command output injection

See the Gemini CLI documentation for more details.