setup-auto-memory

Setting up structured project memory in Claude Code involves installing a per-project memory directory, header index, and documentation fragments automatically. This requires a Memnyx checkout, an active project slug under ~/.claude/projects, and per-step confirmation to protect existing files.

Auto-memory for Claude Code is a structured, persistent memory system that creates per-project memory directories and optional example seeding. It includes a memory-hygiene skill for ongoing curation and a CLAUDE.md fragment documenting the on-disk memory type system, with every write gated by explicit confirmation.

Yes, enabling persistent Claude memory requires an active project slug under the ~/.claude/projects directory. You also need a Memnyx checkout with access to auto-memory and the .claude environment to successfully apply the structured memory setup.

Yes, you can optionally seed sanitized example memory files during workspace memory configuration. These examples demonstrate the correct file shapes and frontmatter structure, making it easier to understand and implement your own persistent memory entries.

Setting up auto-memory protects existing files by requiring per-step confirmation during the installation process. This gating mechanism ensures that the structured memory directory, curation tooling, and CLAUDE.md fragments are applied safely without clobbering your current workspace configuration.

The best way to maintain Claude Code memory hygiene across sessions is by installing the dedicated memory-hygiene skill during the initial auto-memory setup. This tool provides ongoing curation and maintenance capabilities for your structured, persistent project memory directories.