memories-cli

What This Skill Does

The memories-cli skill provides a command-line interface for managing persistent memory for AI agents. It bridges the gap between various AI development environments like Claude, Cursor, and Windsurf by centralizing project rules, architectural decisions, and configuration files. It acts as a single source of truth for your agent's knowledge, ensuring that instructions, preferences, and project-specific conventions are synchronized, searchable, and always accessible during your development lifecycle.

Installation

You can install the skill via the OpenClaw hub or directly via npm to manage your project's memory layer. Run the following command: clawhub install openclaw/skills/skills/charlesrhoward/memories-cli

Once installed, navigate to your project root and execute memories init. This command auto-detects your existing AI tool environment (such as Cursor or Claude) and generates the necessary configuration files to establish a persistent memory link.

Use Cases

• Project Onboarding: Automatically ingest existing .cursorrules or CLAUDE.md files into a centralized database.
• Knowledge Management: Store architectural decisions, API preferences, or "do's and don'ts" as tagged memories to ensure the AI follows them consistently.
• Cross-Tool Syncing: Use memories generate to push your centralized rules to multiple AI tools simultaneously, keeping your development environment in perfect sync.
• Searchable History: Use memories search to recall specific implementation details from past sessions without manually auditing hundreds of lines of code.

Example Prompts

1. "Initialize memories in this directory and import all existing rules from my .cursor/rules folder."
2. "Add a memory that we are using Supabase for authentication and enforce this in all future AI generated code."
3. "Search for all memories tagged with 'api' and generate updated CLAUDE.md and .cursorrules files based on these insights."

Tips & Limitations

• Pro-tip: Use specific types when adding memories (e.g., --type rule vs --type decision) to make your memories list commands more effective.
• Syncing: Always run memories sync after major project changes to ensure your remote cloud database is up-to-date for multi-machine development.
• Limitations: While highly robust, the CLI requires local filesystem access. If you are operating in a highly restricted sandbox or browser-only environment, consider using the memories-mcp server as an alternative interface to the same data store.