Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Diagnose a tangled git situation and give the exact, safe commands to fix it. Use when asked to undo a commit, recover lost work, fix a bad merge or rebase, resolve a detached HEAD, unstage files, or get out of a git mess. Produces the diagnosis, the precise commands to run in order, what each does, and a recovery note if something goes wrong.
.claude/skills/mohitagw15856-git-troubleshooter/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 17% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -9% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 5% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-15 | ✗→✓ | ▲ Improved | -9% | 0% |
Get the user un-stuck from git — calmly, safely, and without destroying work.
Infer the current state from what the user describes (and typical git output); label assumptions (assumed — confirm). Always give a concrete command sequence. If a step is destructive, say so loudly before it.
What happened / what they want (e.g. "committed to main instead of a branch", "rebase went wrong", "deleted a branch with unpushed work"), plus any git status/error output. Infer the rest.
One or two lines: what state the repo is in and why the user is stuck.
A numbered list of exact commands, each with a one-line note of what it does:
1. git reflog # find the lost commit's SHA
2. git checkout -b rescue <SHA> # recover it onto a new branchPrefer non-destructive routes (branch, reflog, --soft) over destructive ones. Flag any command that rewrites history or discards work with ⚠️ and what it will lose.
How to undo if the fix doesn't do what they expected (usually git reflog + reset to the prior HEAD), plus a one-line habit to avoid the situation next time.
git push --force, reset --hard, or clean -fd without a ⚠️ and a safer alternative firstgit reflog when work might be recoverable — it usually is| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-06 | pass→pass | 8,979 | 6,249 | -30% | 1 | 1 | 0% | 1,727 | 1,511 | -13% | 0 | 0 | — |
case-05 | pass→pass | 9,611 | 8,305 | -14% | 1 | 1 | 0% | 1,753 | 1,946 | +11% | 0 | 0 | — |
case-01 | fail→fail | 11,540 | 41,454 | +259% | 1 | 1 | 0% | 2,046 | 2,428 | +19% | 0 | 0 | — |
case-02 | fail→fail | 8,545 | 7,704 | -10% | 1 | 1 | 0% | 1,584 | 1,828 | +15% | 0 | 0 | — |
case-03 | fail→pass | 11,855 | 11,368 | -4% | 1 | 1 | 0% | 2,128 | 2,491 | +17% | 0 | 0 | — |
case-04 | pass→fail | 10,306 | 8,466 | -18% | 1 | 1 | 0% | 1,804 | 1,863 | +3% | 0 | 0 | — |
case-07 | fail→fail | 40,798 | 8,608 | -79% | 1 | 1 | 0% | 1,893 | 1,863 | -2% | 0 | 0 | — |
case-08 | fail→fail | 47,122 | 13,766 | -71% | 1 | 1 | 0% | 2,815 | 2,679 | -5% | 0 | 0 | — |
case-09 | fail→pass | 15,125 | 11,718 | -23% | 1 | 1 | 0% | 2,648 | 2,403 | -9% | 0 | 0 | — |
case-10 | fail→pass | 12,096 | 9,935 | -18% | 1 | 1 | 0% | 1,996 | 2,097 | +5% | 0 | 0 | — |
case-11 | fail→fail | 14,602 | 9,042 | -38% | 1 | 1 | 0% | 2,317 | 1,896 | -18% | 0 | 0 | — |
case-12 | pass→pass | 13,811 | 8,526 | -38% | 1 | 1 | 0% | 2,422 | 1,898 | -22% | 0 | 0 | — |
case-13 | fail→pass | 9,396 | 7,666 | -18% | 1 | 1 | 0% | 1,598 | 1,692 | +6% | 0 | 0 | — |
case-14 | fail→fail | 12,507 | 10,471 | -16% | 1 | 1 | 0% | 2,125 | 2,170 | +2% | 0 | 0 | — |
case-15 | fail→pass | 13,172 | 11,008 | -16% | 1 | 1 | 0% | 2,256 | 2,062 | -9% | 0 | 0 | — |
case-16 | fail→fail | 12,895 | 10,062 | -22% | 1 | 1 | 0% | 2,227 | 2,142 | -4% | 0 | 0 | — |
case-17 | fail→pass | 14,103 | 9,829 | -30% | 1 | 1 | 0% | 2,475 | 2,022 | -18% | 0 | 0 | — |
case-18 | fail→pass | 13,629 | 8,156 | -40% | 1 | 1 | 0% | 2,249 | 1,865 | -17% | 0 | 0 | — |
case-19 | fail→fail | 10,639 | 7,687 | -28% | 1 | 1 | 0% | 1,828 | 1,691 | -7% | 0 | 0 | — |
case-20 | fail→pass | 10,457 | 7,129 | -32% | 1 | 1 | 0% | 1,744 | 1,629 | -7% | 0 | 0 | — |
case-21 | fail→pass | 11,495 | 10,413 | -9% | 1 | 1 | 0% | 2,058 | 2,243 | +9% | 0 | 0 | — |
case-22 | fail→pass | 12,581 | 10,661 | -15% | 1 | 1 | 0% | 2,185 | 2,234 | +2% | 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 +41 percentage points is the difference between those two pass rates over the 22 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/15/2026 | +18% |
Other measured skills in the registry, with their headline benchmark lift.