Does it get more accurate over time?

Yes, it includes a calibration feature where you feed back actual completion times to refine future estimation multipliers.

What is Progressive Estimation?

It is a methodology and Claude Code skill used to estimate software tasks by accounting for the increased velocity provided by AI agents using statistical models.

How does it handle different team structures?

The skill categorizes workflows into human-only, hybrid, or agent-first modes, applying specific research-backed multipliers for each category.

What are P75 and P90 intervals?

These are statistical confidence levels; a P90 estimate means there is a 90% probability that the task will be completed within the suggested timeframe.

Can I use this with Jira or Linear?

Yes, the skill supports output formatting specifically for major project management tools including Jira, Linear, GitHub Issues, and ClickUp.

Progressive Estimation

Name: Progressive Estimation
Author: sickn33

bysickn33

•

31,721

•

Herramientas de Colaboración

Estimates AI-assisted and hybrid software development tasks using research-backed PERT statistics and confidence intervals.

Progressive Estimation is a specialized Claude Code skill designed to bring statistical rigor to sprint planning and backlog management in the age of AI agents. It intelligently detects team working modes—whether human-only, hybrid, or agent-first—and applies empirical velocity multipliers to generate realistic P50 to P90 confidence intervals. By utilizing three-point estimation and a calibration feedback loop, this skill helps teams move away from gut-feeling estimates toward data-driven forecasting for complex software projects involving AI assistance.

Características Principales

01Hybrid mode detection for human and AI-agent collaboration

02Platform-agnostic formatting for Jira, Linear, and GitHub Issues

03Calibration feedback loops to improve accuracy over time

0431,721 GitHub stars

05Confidence band generation including P50, P75, and P90 intervals

06Statistical PERT calculation for three-point work estimation

Casos de Uso

01Sprint planning for teams utilizing AI agents to speed up development

02Batch sizing large backlogs of technical debt or feature requests

03Forecasting release dates with specific confidence intervals for stakeholders

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add sickn33/antigravity-awesome-skills progressive-estimation

For use in Claude.ai and ChatGPT

Características Principales

01Hybrid mode detection for human and AI-agent collaboration

02Platform-agnostic formatting for Jira, Linear, and GitHub Issues

03Calibration feedback loops to improve accuracy over time

0431,721 GitHub stars

05Confidence band generation including P50, P75, and P90 intervals

06Statistical PERT calculation for three-point work estimation

Casos de Uso

01Sprint planning for teams utilizing AI agents to speed up development

02Batch sizing large backlogs of technical debt or feature requests

03Forecasting release dates with specific confidence intervals for stakeholders

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add sickn33/antigravity-awesome-skills progressive-estimation

For use in Claude.ai and ChatGPT