worklog

To generate a weekly worklog automatically, this skill processes per-session JSONL data to extract user requests, touched files, tools used, and assistant conclusions into a structured Markdown file. It aggregates all sessions within the workweek and groups the results by project path for organized archival.

The best way to summarize session logs by project is to group all workweek sessions by their project path. This skill automatically extracts key signals like user requests and touched files from the JSONL data to produce a clear, project-organized activity record in Markdown format.

Extracting JSONL session data for Markdown reports involves reading per-session JSONL files to pull user requests, touched files, tools used, and assistant conclusions. The processed data is then compiled and written into a structured Markdown file named YYYY-WXX under the worklog directory.

No external dependencies are required to automate your Claude Code session summaries. This skill operates independently without any listed dependencies, directly reading the native JSONL session logs to generate the weekly Markdown worklog file.

Yes, you can compile a weekly worklog across multiple project paths. This skill aggregates all sessions from the workweek and specifically groups the extracted results by project path, ensuring that activities from different projects are organized separately in the final Markdown summary.

The specific details extracted from Claude Code logs for the worklog include user requests, touched files, tools used, and assistant conclusions. These key signals are pulled directly from the per-session JSONL data to produce a structured summary of your weekly activity.