sprint-retrospective_skill

Overview

This skill facilitates effective sprint retrospectives to drive continuous team improvement. It guides facilitators through multiple retrospective formats, creates concise action items, and enforces safe, time-boxed sessions. Use it to structure outcomes, track follow-ups, and strengthen team collaboration.

How this skill works

The skill proposes proven retrospective formats (Start-Stop-Continue, Mad-Sad-Glad, 4Ls), collects observations across three core categories (what went well, what didn’t, and actionable improvements), and produces a compact retrospective document with prioritized action items and key metrics. It enforces constraints like a blame-free environment, 2–3 concrete action items, and follow-up checks in the next retro. Facilitators receive facilitation tips, time-box suggestions, and a ready-to-share summary.

When to use it

• At the end of every sprint to reflect on recent work and outcomes
• After a major release or project milestone to capture lessons learned
• When recurring issues or team friction appear and immediate improvement is needed
• To onboard new facilitators and standardize retrospective practices
• When you need a concise artifact for stakeholders summarizing improvements and metrics

Best practices

• Keep the session time-boxed (aim for 45–60 minutes) and stick to the format
• Rotate the facilitator to build shared ownership and varied perspectives
• Limit action items to 2–3 concrete, assigned tasks with due dates
• Maintain a safe, non-blaming environment; focus on processes, not people
• Capture a brief follow-up check for each action item in the next retrospective

Example use cases

• Run a 45-minute Start-Stop-Continue retro at sprint end and export a one-page summary with three action items and key metrics
• Use Mad-Sad-Glad after a problematic deploy to surface emotional responses and immediate fixes (e.g., strengthen a deployment checklist)
• Apply 4Ls after adopting a new tool to document what worked, what was learned, and what’s still missing
• Create a recurring retro template for distributed teams to standardize asynchronous inputs before the live session
• Facilitate a post-mortem style retro after a high-severity incident, focusing on process changes and follow-up owners