Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create a git worktree from a target branch with a new branch name
.claude/skills/nudgebee-worktree/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | -10% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -12% | 0% |
| case-21 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-11 | ✓→✗ | ▼ Worse | -32% | 0% |
| case-22 | ✓→✗ | ▼ Worse | -9% | 0% |
Create a new git worktree checked out from a target branch. Arguments: $ARGUMENTS should be in the format <target-branch> <new-branch-name>.
Examples:
/worktree main feature/add-auth/worktree prod hotfix/fix-crash/worktree test fix/flaky-testsExtract from $ARGUMENTS:
main, test, prod)If arguments are missing, ask the user to provide them in the format: /worktree <target-branch> <new-branch-name>
bashgit fetch origin
Place the worktree as a sibling directory to the current repo:
bash# If repo is at /Users/user/work/nudgebee/nudgebee # Worktree goes to /Users/user/work/nudgebee/nudgebee-<new-branch-name> REPO_ROOT=$(git rev-parse --show-toplevel) WORKTREE_DIR="${REPO_ROOT}-$(echo '<new-branch-name>' | tr '/' '-')"
The branch name's slashes are converted to dashes for the directory name (e.g., feature/add-auth becomes nudgebee-feature-add-auth).
bashgit worktree add -b <new-branch-name> "$WORKTREE_DIR" origin/<target-branch>
This creates a new branch <new-branch-name> based on origin/<target-branch> and checks it out in the worktree directory.
bash# Show all worktrees git worktree list
Output:
Worktree created:
Path: {worktree_dir}
Branch: {new-branch-name}
Based on: origin/{target-branch}
To start working:
cd {worktree_dir}
To remove later:
git worktree remove {worktree_dir}| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 6,152 | 5,748 | -7% | 1 | 1 | 0% | 986 | 948 | -4% | 0 | 0 | — |
case-02 | fail→fail | 5,415 | 5,132 | -5% | 1 | 1 | 0% | 949 | 842 | -11% | 0 | 0 | — |
case-03 | fail→fail | 4,812 | 2,432 | -49% | 1 | 1 | 0% | 792 | 818 | +3% | 0 | 0 | — |
case-04 | pass→pass | 7,131 | 5,086 | -29% | 1 | 1 | 0% | 1,225 | 1,370 | +12% | 0 | 0 | — |
case-05 | pass→pass | 8,910 | 7,453 | -16% | 1 | 1 | 0% | 1,563 | 1,845 | +18% | 0 | 0 | — |
case-06 | pass→pass | 5,139 | 2,497 | -51% | 1 | 1 | 0% | 876 | 958 | +9% | 0 | 0 | — |
case-07 | fail→pass | 6,233 | 2,357 | -62% | 1 | 1 | 0% | 1,041 | 934 | -10% | 0 | 0 | — |
case-08 | fail→pass | 9,741 | 5,261 | -46% | 1 | 1 | 0% | 1,620 | 1,422 | -12% | 0 | 0 | — |
case-09 | pass→pass | 7,237 | 4,855 | -33% | 1 | 1 | 0% | 1,186 | 1,402 | +18% | 0 | 0 | — |
case-10 | fail→fail | 8,774 | 4,822 | -45% | 1 | 1 | 0% | 1,422 | 1,258 | -12% | 0 | 0 | — |
case-11 | pass→fail | 10,014 | 3,989 | -60% | 1 | 1 | 0% | 1,708 | 1,164 | -32% | 0 | 0 | — |
case-12 | pass→pass | 9,705 | 4,148 | -57% | 1 | 1 | 0% | 1,704 | 1,250 | -27% | 0 | 0 | — |
case-13 | pass→pass | 7,033 | 4,054 | -42% | 1 | 1 | 0% | 1,150 | 1,198 | +4% | 0 | 0 | — |
case-14 | pass→pass | 8,597 | 4,098 | -52% | 1 | 1 | 0% | 1,472 | 1,205 | -18% | 0 | 0 | — |
case-15 | fail→fail | 6,994 | 2,385 | -66% | 1 | 1 | 0% | 1,128 | 916 | -19% | 0 | 0 | — |
case-16 | fail→fail | 7,942 | 2,446 | -69% | 1 | 1 | 0% | 1,405 | 976 | -31% | 0 | 0 | — |
case-17 | pass→pass | 9,497 | 3,791 | -60% | 1 | 1 | 0% | 1,695 | 1,232 | -27% | 0 | 0 | — |
case-18 | pass→pass | 9,036 | 3,778 | -58% | 1 | 1 | 0% | 1,588 | 1,183 | -26% | 0 | 0 | — |
case-19 | pass→pass | 9,541 | 3,566 | -63% | 1 | 1 | 0% | 1,576 | 1,171 | -26% | 0 | 0 | — |
case-20 | pass→pass | 5,241 | 3,456 | -34% | 1 | 1 | 0% | 845 | 1,121 | +33% | 0 | 0 | — |
case-21 | fail→pass | 10,296 | 3,688 | -64% | 1 | 1 | 0% | 1,823 | 1,211 | -34% | 0 | 0 | — |
case-22 | pass→fail | 8,089 | 4,289 | -47% | 1 | 1 | 0% | 1,349 | 1,222 | -9% | 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 0 percentage points is the difference between those two pass rates over the 22 comparable cases. 3 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.