Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Automatically advance to the next logical step in the GSD workflow
.claude/skills/coco-research-gsd-next/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-10 | ✗→✓ | ▲ Improved | -65% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -70% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -61% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -76% | 0% |
| case-21 | ✗→✓ | ▲ Improved | -77% | 0% |
<objective> Detect the current project state and automatically invoke the next logical GSD workflow step. No arguments needed — reads STATE.md, ROADMAP.md, and phase directories to determine what comes next.
Designed for rapid multi-project workflows where remembering which phase/step you're on is overhead.
Supports --force flag to bypass safety gates (checkpoint, error state, verification failures). </objective>
<execution_context> @$HOME/.claude/get-shit-done/workflows/next.md </execution_context>
<process> Execute the next workflow from @$HOME/.claude/get-shit-done/workflows/next.md end-to-end. </process>
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 9,663 | 14,591 | +51% | 1 | 1 | 0% | 694 | 430 | -38% | 0 | 0 | — |
case-23 | pass→pass | 17,990 | 8,612 | -52% | 1 | 1 | 0% | 2,414 | 335 | -86% | 0 | 0 | — |
case-02 | fail→fail | 8,612 | 11,442 | +33% | 1 | 1 | 0% | 437 | 441 | +1% | 0 | 0 | — |
case-03 | fail→fail | 8,477 | 11,298 | +33% | 1 | 1 | 0% | 495 | 458 | -7% | 0 | 0 | — |
case-04 | pass→pass | 16,488 | 10,089 | -39% | 1 | 1 | 0% | 1,964 | 935 | -52% | 0 | 0 | — |
case-05 | pass→fail | 23,416 | 15,052 | -36% | 1 | 1 | 0% | 2,922 | 522 | -82% | 0 | 0 | — |
case-06 | pass→fail | 14,295 | 6,524 | -54% | 1 | 1 | 0% | 1,647 | 549 | -67% | 0 | 0 | — |
case-07 | fail→fail | 22,948 | 10,919 | -52% | 1 | 1 | 0% | 1,185 | 1,193 | +1% | 0 | 0 | — |
case-08 | fail→fail | 15,887 | 10,606 | -33% | 1 | 1 | 0% | 1,829 | 330 | -82% | 0 | 0 | — |
case-09 | fail→fail | 14,524 | 3,974 | -73% | 1 | 1 | 0% | 1,484 | 530 | -64% | 0 | 0 | — |
case-10 | fail→pass | 20,131 | 20,256 | +1% | 1 | 1 | 0% | 3,037 | 1,064 | -65% | 0 | 0 | — |
case-11 | fail→pass | 20,630 | 8,385 | -59% | 1 | 1 | 0% | 2,373 | 712 | -70% | 0 | 0 | — |
case-12 | fail→pass | 15,744 | 7,707 | -51% | 1 | 1 | 0% | 1,497 | 587 | -61% | 0 | 0 | — |
case-13 | fail→fail | 14,396 | 8,496 | -41% | 1 | 1 | 0% | 1,452 | 1,289 | -11% | 0 | 0 | — |
case-14 | fail→pass | 13,412 | 1,547 | -88% | 1 | 1 | 0% | 1,458 | 356 | -76% | 0 | 0 | — |
case-15 | pass→pass | 14,993 | 4,319 | -71% | 1 | 1 | 0% | 2,171 | 875 | -60% | 0 | 0 | — |
case-16 | fail→fail | 16,642 | 15,872 | -5% | 1 | 1 | 0% | 1,995 | 467 | -77% | 0 | 0 | — |
case-17 | fail→fail | 13,020 | 18,394 | +41% | 1 | 1 | 0% | 1,611 | 680 | -58% | 0 | 0 | — |
case-18 | pass→pass | 19,253 | 6,772 | -65% | 1 | 1 | 0% | 2,200 | 400 | -82% | 0 | 0 | — |
case-19 | fail→fail | 10,635 | 2,224 | -79% | 1 | 1 | 0% | 1,681 | 477 | -72% | 0 | 0 | — |
case-20 | fail→fail | 10,606 | 13,593 | +28% | 1 | 1 | 0% | 1,646 | 786 | -52% | 0 | 0 | — |
case-21 | fail→pass | 17,254 | 6,951 | -60% | 1 | 1 | 0% | 1,737 | 399 | -77% | 0 | 0 | — |
case-22 | fail→pass | 4,282 | 7,882 | +84% | 1 | 1 | 0% | 591 | 1,512 | +156% | 0 | 0 | — |
case-24 | fail→fail | 16,746 | 11,052 | -34% | 1 | 1 | 0% | 1,987 | 1,170 | -41% | 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. 24 cases were attempted, and 16 counted toward the lift figure. The other 8 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +17 percentage points is the difference between those two pass rates over the 16 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
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.