browser-history_skill

Overview

This skill searches a local browser history database to find visited pages, recover forgotten URLs, and analyze browsing patterns like time spent and most visited sites. It supports both Firefox and Chromium history formats and uses SQLite queries to extract results. Results are formatted for readability, grouped by domain when helpful, and include visit dates and durations.

How this skill works

The skill discovers available browser history files and selects the most recently used browser by default (or asks if multiple were active recently). It opens the appropriate SQLite file in read-only/immutable mode and runs tailored SQL queries for Firefox or Chromium schemas. Queries return titles, URLs, visit timestamps, visit counts, and metadata used to compute time spent and summaries.

When to use it

• You need to find a page you visited but can’t recall the URL or title.
• You want a list of pages visited during a specific date or date range.
• You want browsing statistics: most visited domains, session counts, or time spent on sites.
• You want to reconstruct activity (e.g., what GitHub repos you viewed last week).
• You need a domain-level summary or daily histogram of visits for analysis.

Best practices

• Run the browser-detection step first and default to the most recently used history file.
• Open the SQLite file with an immutable/read-only URI to avoid conflicts while the browser is open.
• Use keyword and date-range filters to limit result size and improve relevance.
• Aggregate by domain for long result sets; show top N domains with totals and an option to expand.
• Convert raw timestamps and millisecond durations to human-friendly formats (local date, hours/minutes).

Example use cases

• Find the article about large language models you read two weeks ago by searching titles and URLs.
• Show the top 20 most visited domains for last month and total time spent per domain.
• List every visited GitHub URL during a specified week to recreate what repositories were inspected.
• Compute total minutes spent on a given site using the browser’s metadata tables (when available).
• Produce a daily histogram of visit counts for the past 30 days to spot browsing spikes.