Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create an isolated workspace with repo copies and independent .planning/
.claude/skills/coco-research-gsd-new-workspace/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-16 | ✗→✓ | ▲ Improved | -58% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -38% | 0% |
| case-19 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-21 | ✗→✓ | ▲ Improved | -71% | 0% |
| case-05 | ✓→✗ | ▼ Worse | 98% | 0% |
<context> Flags:
--name (required) — Workspace name--repos — Comma-separated repo paths or names. If omitted, interactive selection from child git repos in cwd--path — Target directory. Defaults to ~/gsd-workspaces/<name>--strategy — worktree (default, lightweight) or clone (fully independent)--branch — Branch to checkout. Defaults to workspace/<name>--auto — Skip interactive questions, use defaults</context>
<objective> Create a physical workspace directory containing copies of specified git repos (as worktrees or clones) with an independent .planning/ directory for isolated GSD sessions.
Use cases:
.planning/Creates:
<path>/WORKSPACE.md — workspace manifest<path>/.planning/ — independent planning directory<path>/<repo>/ — git worktree or clone for each specified repoAfter this command: cd into the workspace and run /gsd-new-project to initialize GSD. </objective>
<execution_context> @$HOME/.claude/get-shit-done/workflows/new-workspace.md @$HOME/.claude/get-shit-done/references/ui-brand.md </execution_context>
<process> Execute the new-workspace workflow from @$HOME/.claude/get-shit-done/workflows/new-workspace.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-10 | fail→fail | 15,463 | 9,958 | -36% | 1 | 1 | 0% | 280 | 557 | +99% | 0 | 0 | — |
case-16 | fail→pass | 7,605 | 6,533 | -14% | 1 | 1 | 0% | 1,442 | 601 | -58% | 0 | 0 | — |
case-01 | fail→fail | 14,555 | 11,831 | -19% | 1 | 1 | 0% | 2,780 | 713 | -74% | 0 | 0 | — |
case-02 | fail→fail | 16,029 | 10,598 | -34% | 1 | 1 | 0% | 1,729 | 661 | -62% | 0 | 0 | — |
case-03 | fail→fail | 10,541 | 15,725 | +49% | 1 | 1 | 0% | 2,166 | 799 | -63% | 0 | 0 | — |
case-04 | fail→fail | 13,628 | 18,725 | +37% | 1 | 1 | 0% | 1,394 | 813 | -42% | 0 | 0 | — |
case-17 | fail→pass | 8,294 | 7,336 | -12% | 1 | 1 | 0% | 1,230 | 759 | -38% | 0 | 0 | — |
case-05 | pass→fail | 3,535 | 15,988 | +352% | 1 | 1 | 0% | 644 | 1,277 | +98% | 0 | 0 | — |
case-06 | pass→fail | 1,509 | 20,300 | +1245% | 1 | 1 | 0% | 266 | 1,231 | +363% | 0 | 0 | — |
case-07 | fail→fail | 14,906 | 12,742 | -15% | 1 | 1 | 0% | 272 | 767 | +182% | 0 | 0 | — |
case-08 | fail→fail | 12,767 | 6,058 | -53% | 1 | 1 | 0% | 1,359 | 702 | -48% | 0 | 0 | — |
case-09 | fail→fail | 5,629 | 6,284 | +12% | 1 | 1 | 0% | 882 | 841 | -5% | 0 | 0 | — |
case-11 | fail→fail | 6,278 | 11,677 | +86% | 1 | 1 | 0% | 1,144 | 688 | -40% | 0 | 0 | — |
case-12 | fail→fail | 64,387 | 15,335 | -76% | 1 | 1 | 0% | 235 | 705 | +200% | 0 | 0 | — |
case-13 | fail→fail | 12,322 | 11,462 | -7% | 1 | 1 | 0% | 1,388 | 748 | -46% | 0 | 0 | — |
case-14 | pass→fail | 14,142 | 10,695 | -24% | 1 | 1 | 0% | 1,197 | 637 | -47% | 0 | 0 | — |
case-15 | fail→fail | 15,393 | 18,709 | +22% | 1 | 1 | 0% | 2,387 | 1,077 | -55% | 0 | 0 | — |
case-18 | pass→pass | 15,416 | 3,601 | -77% | 1 | 1 | 0% | 1,687 | 865 | -49% | 0 | 0 | — |
case-19 | fail→pass | 10,537 | 6,788 | -36% | 1 | 1 | 0% | 903 | 600 | -34% | 0 | 0 | — |
case-20 | pass→pass | 24,496 | 6,638 | -73% | 1 | 1 | 0% | 3,353 | 625 | -81% | 0 | 0 | — |
case-21 | fail→pass | 21,581 | 6,579 | -70% | 1 | 1 | 0% | 2,355 | 688 | -71% | 0 | 0 | — |
case-22 | pass→pass | 16,078 | 6,859 | -57% | 1 | 1 | 0% | 1,551 | 639 | -59% | 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 7 counted toward the lift figure. The other 15 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 +5 percentage points is the difference between those two pass rates over the 7 comparable cases. 6 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.