Coding Wrapped Makes Vibe Coding Visible

The clever part isn’t the pixel art—it’s making AI-assisted coding habits visible without turning personal development history into cloud analytics.

• –Scans standard Claude Code and Codex session folders, then excludes raw transcripts, source code, project names, paths, URLs, emails, and secrets from generated state.
• –Combines factual metrics with four illustrated insights covering user behavior, agent behavior, working style, and practical next steps.
• –The shared Agent Skill works across Claude Code and Codex through one installation path, reducing platform-specific drift.
• –Its “fun before judgment” framing is more inviting than productivity scoring, which could make developers more willing to inspect and improve their workflows.
• –The main limitation is scope: it is a personal retrospective tool, not team analytics, productivity measurement, or broad multi-agent observability.