This turns ~/.memory/ into a persistent knowledge store that survives across projects and conversations. It auto-loads relevant memories at conversation start by matching tags against your current context, caps each file at 100 lines to stay context-efficient, and uses wiki-links to cross-reference entries. The smart bit is that it asks before saving anything and proposes what to remember after complex tasks instead of silently hoarding everything. Best for people who work across multiple projects and want the AI to remember their preferences, workflows, and past learnings without cluttering project-specific context files. The structured onboarding and tag-based relevance matching make it more surgical than dumping everything into a single context file.
npx -y skills add absolutelyskilled/absolutelyskilled --skill second-brain --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
juliusbrussee/caveman
mattpocock/skills
shadcn/improve
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills