What problem does it solve? Claude Code stores every session as scattered .jsonl files across project directories, making it hard to recall what you worked on, when, and in which repository. This Skill mines those logs to reconstruct a chronological timeline of your past sessions without any manual searching. ## Core Features & Use Cases - Session Timeline Mining: Scans Claude Code project directories (including worktrees) and lists the N most recent sessions with start/end times, duration, message counts, and summaries. - Gap Detection: Automatically inserts gap markers between sessions separated by more than 30 minutes, showing when you were offline or idle. - Repo Attribution: Resolves each session to a human-readable repository name using ghq, and supports a day-by-day grouped timeline view across all repos. - Use Case: You return Monday morning and ask "what did I work on last week" — the Skill renders a day-grouped timeline of sessions, durations, and focus summaries across every repo. ## Quick Start Ask the assistant to run /dig to show my recent Claude Code sessions as a timeline with gaps and repo names.
