harness-retrospective

Conduct structured post-epic retrospectives with evidence-based analysis and next-epic preparation.

1|Updated May 23, 2026
One-click install
npx skills add https://github.com/baobao0303/harness --skill harness-retrospective
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: harness-retrospective
Source: https://github.com/baobao0303/harness/tree/main/.agents/skills/harness-retrospective
Command: npx skills add https://github.com/baobao0303/harness --skill harness-retrospective

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps teams produce a blame-free, evidence-based retrospective after an epic so lessons, continuity, and next-epic preparation become concrete action items.

Core Features & Use Cases

  • Epic selection with validation: Detects the most relevant completed epic using sprint status, with fallbacks to user-provided selection and story-folder inference when needed.
  • Evidence-driven analysis: Loads epic PRD/architecture context and synthesizes story-record patterns (dev notes, review feedback, lessons learned, technical debt, and testing signals).
  • Two-part facilitation: Runs an Epic Review (wins/challenges/lessons) and then prepares Next Epic Preparation (dependencies, risks, and actionable commitments).

Quick Start

Ask an AI agent running this Skill to run a retrospective for the most recently completed epic and generate the retrospective analysis plus next epic preparation with assigned action items.

Frequently Asked Questions about harness-retrospective

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

FAQPage Schema
How do I conduct a structured epic retrospective after a sprint?

An epic retrospective analyzes completed sprint stories and project context to extract lessons, validate epic completion, and generate actionable next-epic readiness commitments. It synthesizes story records, dev notes, and architecture documents to produce evidence-based action items.

What is the best way to track technical debt and lessons learned from a completed epic?

Tracking technical debt and lessons learned requires synthesizing story-record patterns and review feedback from completed epics. This process extracts testing signals and dev notes to validate epic completion, preparing concrete action items for the next development phase.

Can I automate blame-free post-epic reviews using sprint status signals?

Yes, blame-free post-epic reviews can be automated by enforcing psychological safety constraints and loading sprint status signals. The system detects the most relevant completed epic and synthesizes evidence from story folders to generate objective lessons.

How do I prepare next epic readiness and dependencies from previous sprint data?

Next epic readiness is prepared by running a two-part facilitation process that first reviews epic wins and challenges, then extracts dependencies, risks, and actionable commitments from previous sprint story records and PRD context.

Does a structured epic review work without explicit PRD and architecture documents?

Epic review can function without explicit PRD documents by falling back to user-provided selection and story-folder inference. However, loading architecture context and PRD documents enriches the evidence-driven analysis for validating epic completion.

Why use a systematic retrospective process instead of a standard sprint review?

A systematic retrospective process ensures deterministic facilitation by enforcing psychological safety and requiring two-part output for epic review and next epic preparation. This validates epic completion using story-record evidence rather than subjective opinions.