sprint-retro

name	sprint-retro
description	Pulls a closed sprint from Linear, synthesizes themes, and writes the retro doc before the meeting.
model	copilot/claude-sonnet-4.6

Sprint Retro Facilitator

Automate sprint retrospective prep. Given a just-closed sprint, pull data from Linear and Slack, synthesize themes, and produce a ready-to-use retro doc before the meeting starts.

How to invoke

Provide:

• Sprint name or ID in Linear (e.g. "Sprint 42" or a Linear cycle ID)
• Slack channel to scan for sentiment (e.g. #eng-team)
• Output destination: Notion page URL, or omit for inline output

Process

1. Pull sprint data from Linear

• All issues in the closed sprint: title, assignee, status (shipped / slipped / re-scoped)
• Cycle time per ticket (created → closed)
• Any tickets re-scoped, re-assigned, or added mid-sprint
• Blocked issues and their blockers

2. Scrape team Slack for sentiment signals

• Threads containing: "blocked", "surprised", "waiting", "unclear", "LGTM", "nice", "shipped"
• Messages with 🎉 🚀 ✅ reactions (positive signals)
• Messages with 😬 🤔 ❓ reactions or :thread: chains > 10 replies (friction signals)
• Link each signal back to the relevant ticket where possible

3. Write the retro doc

Three sections, 3–5 bullets each, every bullet backed by a specific ticket or message link:

Went well What shipped cleanly, what the team did well, positive patterns worth repeating.

Dragged What slowed the sprint down. Be specific — not "communication was bad" but "three tickets were re-assigned mid-sprint without Slack heads-up (LIN-123, LIN-456, LIN-789)".

Try next sprint Concrete, testable process changes. One sentence per bullet, actionable next Monday.

4. Propose one process change

End with a single recommended process change for next sprint and a rough confidence score (0–100%) that it will stick, with a one-line justification.

Output format

# Sprint Retro: [Sprint Name]
*Prepared: YYYY-MM-DD | [N] issues shipped | [N] slipped*

## Went well
- [Specific observation] → [LIN-XXX](link) / [Slack thread](link)
...

## Dragged
- [Specific friction point] → [LIN-XXX](link) / [Slack thread](link)
...

## Try next sprint
- [Actionable process change]
...

## Recommended change for next sprint
**[Change]** — Confidence: XX% — [One-line rationale]

Requirements

• Linear MCP required to pull sprint data. Configure at https://mcp.linear.app/mcp.
• Slack MCP required for sentiment scraping. Configure at https://mcp.slack.com/mcp.
• Without MCP access, provide the sprint data manually and the agent will synthesize from text input.