focus-forensics

Analyzes weekly screen-capture snapshots to report context-switching structure and week-over-week change.

84|37|Updated May 23, 2026
One-click install
npx skills add https://github.com/legioncodeinc/vibe-coding-tools --skill focus-forensics-legioncodeinc
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: focus-forensics
Source: https://github.com/legioncodeinc/vibe-coding-tools/tree/main/src/quarantine/skills/keep/focus-forensics
Command: npx skills add https://github.com/legioncodeinc/vibe-coding-tools --skill focus-forensics-legioncodeinc

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? When a week feels scattered, most tools invent numbers like hours lost to context switching that periodic screen capture cannot actually measure. This Skill reports only what the Littlebird snapshot instrument genuinely supports: observed switches between work contexts, run lengths, fragmentation by hour, meeting load, and an honest week-over-week comparison. ## Core Features & Use Cases - Switch and run analysis: Counts observed transitions between work contexts, builds run-length distributions, and flags dense transition bursts using a deterministic Python script. - Week-over-week comparison: Reads its own past routine reports and compares this week against last week under four comparability gates, since same-method comparison is more defensible than any absolute number. - Rabbit hole detection with strict evidence bars: Names at most two recurring topics that appeared alongside the user's own stated intentions, with quoted receipts, and refuses to judge without one. - Use Case: On Monday morning, a scheduled routine reviews the week just ended, reporting that runs of length 1 rose from 41 to 58, naming which hours fragmented most, and suggesting one concrete environmental change to try. ## Quick Start Ask the assistant to run a focus report on my last seven days using my Littlebird capture and show what changed since last week.

Frequently Asked Questions about focus-forensics

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I measure context switching from screen capture data?▼

Label each snapshot with a work context, sort by timestamp, then count adjacent pairs whose labels differ as switches. This Skill automates that pipeline with day-scoped Littlebird MCP queries and a deterministic Python script that computes transition rates and run distributions.

How to run a weekly focus report with Littlebird MCP?▼

Ask for a focus report in Claude Cowork or Claude Code with the Littlebird MCP connected. The Skill retrieves snapshots per day, computes switching metrics, compares against past routine reports, and can set up a recurring Monday 08:00 routine after your approval.

Can screen snapshots measure hours lost to distraction?▼

No. Periodic snapshots observe nothing between captures, so hours lost, percentages of the day, and productivity scores are not measurable. This Skill reports switch counts, run lengths in snapshots, and week-over-week change instead, and explains why when asked for hours.

What are the requirements for using Littlebird focus analysis?▼

You need the Littlebird MCP connected on a Power or Pro plan, plus at least three days with snapshots or 50 total snapshots in the window for the switching analysis. Thinner coverage yields only the meeting-load section.

Can I run attention analysis on a colleague or employee?▼

No. The Skill analyzes only the user's own capture for the user and declines any request targeting another person, including de-identified variants. It offers instead to help you ask that person to run it on themselves.