monthly-retro

Generates a monthly retrospective markdown file from closed stories, ADRs, bugs, and monitor data.

Updated May 25, 2026
One-click install
npx skills add https://github.com/DATA-AI-XYZ/Tandem --skill monthly-retro-data-ai-xyz
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: monthly-retro
Source: https://github.com/DATA-AI-XYZ/Tandem/tree/main/skills/monthly-retro
Command: npx skills add https://github.com/DATA-AI-XYZ/Tandem --skill monthly-retro-data-ai-xyz

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Closing out a month of project work requires manually gathering shipped stories, bug counts, decisions, and metrics scattered across the repository, then writing them up consistently. This Skill automates that assembly into a single structured retro file so no month's review is skipped or reconstructed from memory. ## Core Features & Use Cases - Auto-detected retro window: Resolves the most recently completed full month from the system clock, with an argument override for explicit months. - Derived metrics and shipped lists: Runs bundled scripts to compute stories shipped, bugs filed/fixed, ADRs created, and a grouped What Shipped section directly from the story board. - Template-verbatim output: Writes RETRO-YYYY-MM.md from the kit's retro template with correct frontmatter, an auto-drafted objective Summary, and an honest empty-month state. - Use Case: At the start of June, invoke the retro command to produce RETRO-2026-05.md with metrics, a strategic OKR drift check, and candidate One Change actions for the founder to pick from. ## Quick Start Ask the assistant to run the monthly retrospective for last month and save it to the retros folder.

Frequently Asked Questions about monthly-retro

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

FAQPage Schema
How do I run a monthly retrospective for my project?

Invoke the monthly retro command and it auto-detects the most recently completed full month, gathers closed stories, bugs, ADRs, and monitor history for that window, then drafts RETRO-YYYY-MM.md from the retro template for your review before saving.

How are the retro metrics and shipped stories calculated?

Metrics are derived by scripts that scan the stories directory: a story counts as shipped when its status is done and its completed_at date falls in the target month. The same derivation feeds both the What Shipped list and the Metrics section so they always agree.

Can I run a retro for a specific past month instead of last month?

Yes, pass an explicit month argument such as 2026-03 or March and it overrides the clock-based auto-detection. Without an argument, the skill always targets the most recently completed full month, never the in-progress one.

What happens if nothing shipped during the month?

An empty month still produces a valid retro file with zeroed metrics, a What Shipped section stating nothing shipped, and an honest Summary noting the quiet period. The skill never skips the file or falls back to a busier month.

Why does the retro not report average story cycle time?

Cycle time from completed_at minus started_at was retired because orchestrated execution stamps both timestamps minutes apart in the same run, measuring agent execution rather than lead time. The retro reports difficulty signals like rework rate and first-pass close rate instead.