This gives Claude a persistent knowledge layer across sessions using a local SQLite database. Instead of re-learning your project conventions or asking the same questions twice, it searches stored memories before diving into work and saves verified patterns, fixes, and decisions after confirming they're reusable. The quality gate is strict: no logs, no secrets, no speculation, only knowledge backed by code or explicit instructions. It's scoped to repos or projects, not dumped into a global pile. If you're tired of Claude forgetting how your tests run or which API quirks matter, this turns episodic conversations into something closer to actual working memory. The workflow is search first, store after verification, update when stale.
npx -y skills add codeaholicguy/ai-devkit --skill memory --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
sickn33/antigravity-awesome-skills
moizibnyousaf/ai-agent-skills
github/awesome-copilot