retro

Run a retrospective analyzing mistakes, dead ends, and root causes from completed work.

Updated Aug 7, 2026
One-click install
npx skills add https://github.com/ciaran-slow/MellowMaker --skill retro-ciaran-slow
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: retro
Source: https://github.com/ciaran-slow/MellowMaker/tree/main/.claude/skills/retro
Command: npx skills add https://github.com/ciaran-slow/MellowMaker --skill retro-ciaran-slow

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? After finishing a task or issue, teams often move on without understanding what went wrong, what wasted effort, or why failures repeated. This Skill structures a retrospective over the chat history, issue cycle, and workflow so root causes surface and durable fixes get routed into skills, docs, or scripts instead of being forgotten. ## Core Features & Use Cases - Root Cause Analysis: Goes beyond surface symptoms to identify underlying causes in prompts, skills, tools, or workflow sequencing. - Issue-Cycle Review: Closes the plan → build → verify → retro workflow by reviewing the issue, PR, verify findings, and non-blocking gaps. - Durable Fix Routing: Converts learnings into concrete changes to SKILL.md files, repo runbooks, or scripts — never memory. - Use Case: After a PR merges, run the retro to review verify's non-blocking findings, post any missing artifacts to the issue, and commit a test-debt row or doc fix as the retro follow-up. ## Quick Start Ask the assistant to run a retro on the issue that just closed, reviewing the plan, build, and verify stages for mistakes and improvements.

Frequently Asked Questions about retro

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

FAQPage Schema
How do I run a retrospective on a completed task?

Invoke the retro skill after the work finishes. It reviews the chat history, instructions, tools used, and decision sequence, then produces a report covering what went wrong, root causes, what worked, and recommended changes to skills, docs, or scripts.

When should the retro skill run in an issue workflow?

Retro is the mandatory closing stage of the plan → build → verify → retro cycle. Run it on every issue after verify posts its verdict and the PR merges, even when the cycle went smoothly, since clean runs still reveal what to preserve.

What happens to verify's non-blocking findings during a retro?

The retro treats verify's non-blocking findings as its inbox. Each item is either fixed in the retro commit, filed as an issue, or explicitly justified as needing neither, so noted findings do not silently persist across cycles.

Can retro learnings be saved to memory?

No. The skill explicitly forbids saving retro outcomes to memory. Valid destinations are only updated or new SKILL.md files, markdown runbooks or checklists committed to the repo, and deterministic helper scripts.

Does the retro skill always run in one shot?

No. If causes are ambiguous, contested, or benefit from operator insight, it uses an iterative mode: asking a few focused questions about what the user noticed, then refining the root-cause analysis before producing the final retro.