Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Facilitate a structured sprint retrospective — what went well, what didn't, and prioritized action items with owners and deadlines. Use when running a retrospective, reflecting on a sprint, creating action items from team feedback, or learning how to run effective retros.
.claude/skills/phuryn-retro/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 23% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 68% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -5% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 61% | 0% |
Run a structured retrospective that surfaces insights and produces actionable improvements.
You are facilitating a retrospective for $ARGUMENTS.
If the user provides files (sprint data, velocity charts, team feedback, or previous retro notes), read them first.
Format A — Start / Stop / Continue:
Format B — 4Ls (Liked / Learned / Lacked / Longed For):
Format C — Sailboat:
| Priority | Action Item | Owner | Deadline | Success Metric | |---|---|---|---|---| | 1 | Specific, actionable improvement] | Name/Role] | Date] | How we'll know it worked] |
## Sprint X] Retrospective — Date]
### Sprint Performance
### Key Themes
### Action Items
### Carry-over from Last Retro
Save as markdown. Keep the tone constructive — the goal is improvement, not blame.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | pass→pass | 6,506 | 6,464 | -1% | 1 | 1 | 0% | 1,220 | 1,692 | +39% | 0 | 0 | — |
case-01 | fail→pass | 17,127 | 29,309 | +71% | 1 | 1 | 0% | 2,749 | 3,391 | +23% | 0 | 0 | — |
case-02 | fail→pass | 7,383 | 9,079 | +23% | 1 | 1 | 0% | 1,284 | 2,151 | +68% | 0 | 0 | — |
case-03 | fail→pass | 20,632 | 16,420 | -20% | 1 | 1 | 0% | 3,281 | 3,124 | -5% | 0 | 0 | — |
case-05 | pass→pass | 18,331 | 13,721 | -25% | 1 | 1 | 0% | 2,871 | 3,251 | +13% | 0 | 0 | — |
case-06 | pass→pass | 10,258 | 11,489 | +12% | 1 | 1 | 0% | 1,941 | 2,348 | +21% | 0 | 0 | — |
case-07 | pass→pass | 15,888 | 13,640 | -14% | 1 | 1 | 0% | 2,576 | 3,061 | +19% | 0 | 0 | — |
case-08 | fail→pass | 15,923 | 15,250 | -4% | 1 | 1 | 0% | 2,593 | 3,216 | +24% | 0 | 0 | — |
case-09 | fail→fail | 13,228 | 11,209 | -15% | 1 | 1 | 0% | 1,711 | 2,574 | +50% | 0 | 0 | — |
case-10 | fail→fail | 7,221 | 6,371 | -12% | 1 | 1 | 0% | 1,019 | 1,619 | +59% | 0 | 0 | — |
case-11 | fail→pass | 7,957 | 8,151 | +2% | 1 | 1 | 0% | 1,145 | 1,838 | +61% | 0 | 0 | — |
case-12 | fail→pass | 11,831 | 8,671 | -27% | 1 | 1 | 0% | 1,950 | 2,117 | +9% | 0 | 0 | — |
case-13 | pass→pass | 9,853 | 10,136 | +3% | 1 | 1 | 0% | 1,461 | 2,480 | +70% | 0 | 0 | — |
case-14 | pass→pass | 10,294 | 9,472 | -8% | 1 | 1 | 0% | 1,858 | 2,282 | +23% | 0 | 0 | — |
case-15 | fail→fail | 4,355 | 7,638 | +75% | 1 | 1 | 0% | 800 | 2,082 | +160% | 0 | 0 | — |
case-16 | pass→pass | 14,849 | 16,448 | +11% | 1 | 1 | 0% | 2,206 | 2,734 | +24% | 0 | 0 | — |
case-17 | fail→pass | 3,024 | 10,639 | +252% | 1 | 1 | 0% | 463 | 2,048 | +342% | 0 | 0 | — |
case-18 | pass→pass | 14,919 | 12,442 | -17% | 1 | 1 | 0% | 2,332 | 2,527 | +8% | 0 | 0 | — |
case-19 | fail→fail | 12,467 | 8,437 | -32% | 1 | 1 | 0% | 1,803 | 2,006 | +11% | 0 | 0 | — |
case-20 | pass→pass | 9,347 | 9,124 | -2% | 1 | 1 | 0% | 1,505 | 2,244 | +49% | 0 | 0 | — |
case-21 | pass→pass | 11,640 | 13,897 | +19% | 1 | 1 | 0% | 1,957 | 2,370 | +21% | 0 | 0 | — |
case-22 | fail→pass | 10,490 | 9,357 | -11% | 1 | 1 | 0% | 1,864 | 2,342 | +26% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +36 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.