Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Initialize a new project with deep context gathering and PROJECT.md
.claude/skills/coco-research-gsd-new-project/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -65% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -75% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -49% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -3% | 0% |
<runtime_note> Copilot (VS Code): Use vscode_askquestions wherever this workflow calls AskUserQuestion. They are equivalent — vscode_askquestions is the VS Code Copilot implementation of the same interactive question API. </runtime_note>
<context> Flags:
--auto — Automatic mode. After config questions, runs research → requirements → roadmap without further interaction. Expects idea document via @ reference.</context>
<objective> Initialize a new project through unified flow: questioning → research (optional) → requirements → roadmap.
Creates:
.planning/PROJECT.md — project context.planning/config.json — workflow preferences.planning/research/ — domain research (optional).planning/REQUIREMENTS.md — scoped requirements.planning/ROADMAP.md — phase structure.planning/STATE.md — project memoryAfter this command: Run /gsd-plan-phase 1 to start execution. </objective>
<execution_context> @$HOME/.claude/get-shit-done/workflows/new-project.md @$HOME/.claude/get-shit-done/references/questioning.md @$HOME/.claude/get-shit-done/references/ui-brand.md @$HOME/.claude/get-shit-done/templates/project.md @$HOME/.claude/get-shit-done/templates/requirements.md </execution_context>
<process> Execute the new-project workflow from @$HOME/.claude/get-shit-done/workflows/new-project.md end-to-end. Preserve all workflow gates (validation, approvals, commits, routing). </process>
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-21 | fail→fail | 10,983 | 10,168 | -7% | 1 | 1 | 0% | 1,695 | 1,182 | -30% | 0 | 0 | — |
case-01 | fail→fail | 12,800 | 17,156 | +34% | 1 | 1 | 0% | 2,237 | 826 | -63% | 0 | 0 | — |
case-02 | fail→fail | 11,300 | 16,825 | +49% | 1 | 1 | 0% | 311 | 736 | +137% | 0 | 0 | — |
case-03 | fail→fail | 9,262 | 17,430 | +88% | 1 | 1 | 0% | 1,439 | 781 | -46% | 0 | 0 | — |
case-04 | pass→fail | 18,901 | 13,436 | -29% | 1 | 1 | 0% | 4,139 | 958 | -77% | 0 | 0 | — |
case-05 | pass→pass | 9,990 | 14,280 | +43% | 1 | 1 | 0% | 1,409 | 2,038 | +45% | 0 | 0 | — |
case-06 | pass→pass | 22,535 | 17,821 | -21% | 1 | 1 | 0% | 2,584 | 2,672 | +3% | 0 | 0 | — |
case-07 | pass→pass | 12,134 | 10,744 | -11% | 1 | 1 | 0% | 1,837 | 2,021 | +10% | 0 | 0 | — |
case-08 | fail→pass | 11,902 | 2,225 | -81% | 1 | 1 | 0% | 1,226 | 687 | -44% | 0 | 0 | — |
case-09 | fail→pass | 15,437 | 6,850 | -56% | 1 | 1 | 0% | 1,932 | 674 | -65% | 0 | 0 | — |
case-10 | fail→pass | 16,938 | 7,138 | -58% | 1 | 1 | 0% | 2,659 | 669 | -75% | 0 | 0 | — |
case-11 | fail→pass | 12,323 | 7,854 | -36% | 1 | 1 | 0% | 1,720 | 884 | -49% | 0 | 0 | — |
case-12 | fail→pass | 17,936 | 15,252 | -15% | 1 | 1 | 0% | 2,318 | 2,253 | -3% | 0 | 0 | — |
case-13 | fail→fail | 17,984 | 3,601 | -80% | 1 | 1 | 0% | 2,514 | 829 | -67% | 0 | 0 | — |
case-14 | fail→pass | 8,641 | 2,613 | -70% | 1 | 1 | 0% | 1,463 | 674 | -54% | 0 | 0 | — |
case-15 | fail→fail | 34,029 | 1,856 | -95% | 1 | 1 | 0% | 1,687 | 675 | -60% | 0 | 0 | — |
case-16 | fail→pass | 15,192 | 2,454 | -84% | 1 | 1 | 0% | 1,784 | 778 | -56% | 0 | 0 | — |
case-17 | fail→fail | 13,024 | 1,839 | -86% | 1 | 1 | 0% | 1,284 | 673 | -48% | 0 | 0 | — |
case-18 | fail→fail | 14,928 | 7,586 | -49% | 1 | 1 | 0% | 1,411 | 789 | -44% | 0 | 0 | — |
case-19 | fail→fail | 7,512 | 8,588 | +14% | 1 | 1 | 0% | 1,209 | 893 | -26% | 0 | 0 | — |
case-20 | fail→fail | 9,749 | 7,689 | -21% | 1 | 1 | 0% | 1,560 | 857 | -45% | 0 | 0 | — |
case-22 | pass→pass | 15,995 | 16,012 | +0% | 1 | 1 | 0% | 2,505 | 2,090 | -17% | 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, and 18 counted toward the lift figure. The other 4 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 +27 percentage points is the difference between those two pass rates over the 18 comparable cases. 1 case got worse with the skill loaded, and it is 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.