Where does the skill look for session history?

By default, the skill auto-detects the .specstory/history directory within your current project path.

What is 'yak shaving' in the context of this skill?

Yak shaving refers to a coding session that starts with a simple goal (Goal A) but derails into a series of increasingly unrelated tasks (Task Z), such as starting a CSS fix and ending up rewriting a CI/CD pipeline.

What does a score of 100 represent?

A score of 100 represents an 'Epic Rabbit Hole' where the final output has zero relation to the initial user intent and involved multiple major domain shifts.

Can I use this skill to export data for other tools?

Yes, you can use the --json flag to output analysis results in a structured format suitable for further data processing or custom dashboards.

How is the Yak Shave Score calculated?

The score (0-100) is a weighted metric based on domain shifts (40%), goal completion (25%), session length ratio (20%), and tool type escalation (15%).

SpecStory Yak Shave Analyzer

Name: SpecStory Yak Shave Analyzer
Author: organvm-iv-taxis

byorganvm-iv-taxis

•

Productivity & Workflow

Analyzes AI coding session history to detect scope creep and technical rabbit holes.

This skill parses your .specstory/history files to evaluate how well you adhered to your original coding goals during AI-assisted sessions. By tracking domain shifts, tool call patterns, and file references, it generates a 'yak shave score' that highlights when minor tasks evolved into major distractions. It provides personalized LLM-generated insights into your productivity habits, identifying which types of tasks or specific hours of the day lead to the most significant detours, helping you maintain focus and discipline in your development workflow.

Key Features

01Detailed session summaries identifying 'worst offender' rabbit holes

02Proprietary 0-100 scoring algorithm based on domain shifts and goal completion

03Automated parsing of .specstory history files with custom date range filtering

042 GitHub stars

05Trend detection for task types (e.g., UI vs. Backend) and peak drift hours

06Multi-format reporting with support for Markdown and JSON outputs

Use Cases

01Identifying 'danger zone' task categories that typically lead to excessive scope creep

02Reviewing weekly developer productivity to identify and minimize frequent distractions

03Auditing complex refactoring sessions to see where initial intent diverged from execution

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add organvm-iv-taxis/a-i--skills specstory-yak

For use in Claude.ai and ChatGPT

Download Skill