# Project Retrospective

> Project Retrospective

- Skill: `sethdford/project-retrospective` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add sethdford/project-retrospective`
- Raw SKILL.md: https://api.skillmd.com/api/skills/sethdford/project-retrospective/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: sethdford (https://skillmd.com/u/sethdford)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/sethdford/project-retrospective

---

# Project Retrospective

Run a **blameless**, time-boxed retrospective that turns experience into **2–3 concrete actions**—not a wish list. The retro fails if nothing changes; systems and processes are the focus, not individuals.


## When to Use
- End of phase, release, incident follow-up, or recurring team improvement cadence (e.g. monthly).

## Workflow
**Before:** Review status of **last retro’s action items**; drop or reschedule stuck items explicitly. **Time-box:** ~**60 minutes**. **Formats (pick one):** **Start / Stop / Continue**; **4Ls** (Liked, Learned, Lacked, Longed for); **Sailboat** (wind = helps, anchor = hinders, rocks = risks). Facilitate so everyone speaks; capture themes, not vent dumps. **End:** Agree **2–3** specific, owned, dated actions. **After:** Track actions like any other work—the next retro opens with them.

## Examples
**Example 1:** Post-release retro uses 4Ls; team spots “Lacked: staging parity.” Action: add one checklist item to release runbook and owner by next sprint.

**Example 2:** Incident retro uses sailboat; “rocks” include on-call handoff gaps. Actions: update paging policy, run a dry-run drill, assign owners—no naming who “messed up.”