Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Git-aware undo by logical work unit (track, phase, or task)
.claude/skills/conductor-revert/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-19 | ✗→✓ | ▲ Improved | — | — |
| case-07 | ✗→✓ | ▲ Improved | — | — |
| case-20 | ✗→✓ | ▲ Improved | — | — |
| case-15 | ✗→✓ | ▲ Improved | — | — |
| case-09 | ✗→✓ | ▲ Improved | — | — |
Revert changes by logical work unit with full git awareness. Supports reverting entire tracks, specific phases, or individual tasks.
resources/implementation-playbook.md.conductor/tracks.md exists/conductor:setup firstgit status to confirm git repo WARNING: Uncommitted changes detected
Files with changes: {list of files}
Options:
Parse the argument format:
Full track: {trackId}
auth_20250115Specific phase: {trackId}:phase{N}
auth_20250115:phase2Specific task: {trackId}:task{X.Y}
auth_20250115:task2.3Display guided selection menu:
What would you like to revert?
Currently In Progress:
1. [~] Task 2.3 in dashboard_20250112 (most recent)
Recently Completed:
2. [x] Task 2.2 in dashboard_20250112 (1 hour ago)
3. [x] Phase 1 in dashboard_20250112 (3 hours ago)
4. [x] Full track: auth_20250115 (yesterday)
Options:
5. Enter specific reference (track:phase or track:task)
6. Cancel
Select option:bash git log --oneline --grep="{trackId}" --grep="Task {X.Y}" --all-match
bash git log --oneline --grep="mark task {X.Y} complete" --grep="{trackId}" --all-match
bash git log --oneline --grep="{trackId}" | grep -E "Task {N}\.[0-9]"
bash git log --oneline --grep="{trackId}"
bash git log --oneline -- "conductor/tracks/{trackId}/"
Before any revert operations, display full plan:
================================================================================
REVERT EXECUTION PLAN
================================================================================
Target: {description of what's being reverted}
Commits to revert (in reverse chronological order):
1. abc1234 - feat: add chart rendering (dashboard_20250112)
2. def5678 - chore: mark task 2.3 complete (dashboard_20250112)
3. ghi9012 - feat: add data hooks (dashboard_20250112)
4. jkl3456 - chore: mark task 2.2 complete (dashboard_20250112)
Files that will be affected:
- src/components/Dashboard.tsx (modified)
- src/hooks/useData.ts (will be deleted - was created in these commits)
- conductor/tracks/dashboard_20250112/plan.md (modified)
Plan updates:
- Task 2.2: [x] -> [ ]
- Task 2.3: [~] -> [ ]
================================================================================
!! WARNING !!
================================================================================
This operation will:
- Create {N} revert commits
- Modify {M} files
- Reset {P} tasks to pending status
This CANNOT be easily undone without manual intervention.
================================================================================
Type 'YES' to proceed, or anything else to cancel:CRITICAL: Require explicit 'YES' confirmation. Do not proceed on 'y', 'yes', or enter.
Execute reverts in reverse chronological order (newest first):
Executing revert plan...
[1/4] Reverting abc1234...
git revert --no-edit abc1234
✓ Success
[2/4] Reverting def5678...
git revert --no-edit def5678
✓ Success
[3/4] Reverting ghi9012...
git revert --no-edit ghi9012
✓ Success
[4/4] Reverting jkl3456...
git revert --no-edit jkl3456
✓ SuccessIf any revert produces a merge conflict:
================================================================================
MERGE CONFLICT DETECTED
================================================================================
Conflict occurred while reverting: {sha} - {message}
Conflicted files:
- src/components/Dashboard.tsx
Options:
1. Show conflict details
2. Abort revert sequence (keeps completed reverts)
3. Open manual resolution guide
IMPORTANT: Reverts 1-{N} have been completed. You may need to manually
resolve this conflict before continuing or fully undo the revert sequence.
Select option:HALT immediately on any conflict. Do not attempt automatic resolution.
After successful git reverts, update plan.md:
[x] -> [ ][~] -> [ ]tasks.completedstatus if neededupdated timestampDo NOT commit plan.md changes - they are part of the revert operation
[x] or [~] to [ ][~] if partially complete, [ ] if fully revertedAfter revert completion:
================================================================================
REVERT COMPLETE
================================================================================
Summary:
- Reverted {N} commits
- Reset {P} tasks to pending
- {M} files affected
Git log now shows:
{recent commit history}
Plan.md status:
- Task 2.2: [ ] Pending
- Task 2.3: [ ] Pending
================================================================================
Verify the revert was successful:
1. Run tests: {test command}
2. Check application: {relevant check}
If issues are found, you may need to:
- Fix conflicts manually
- Re-implement the reverted tasks
- Use 'git revert HEAD~{N}..HEAD' to undo the reverts
================================================================================git reset --hard - Only use git revertgit push --force - Only safe push operationsNo commits found for track: {trackId}
The track exists but has no associated commits. This may mean:
- Implementation never started
- Commits used different format
Options:
1. Delete track directory only
2. CancelSome commits appear to already be reverted:
- abc1234 was reverted by xyz9876
Options:
1. Skip already-reverted commits
2. Cancel and investigateWARNING: Some commits have been pushed to remote
Commits on remote:
- abc1234 (origin/main)
- def5678 (origin/main)
Reverting will create new revert commits that you'll need to push.
This is the safe approach (no force push required).
Continue with revert? (YES/no):If user needs to undo the revert itself:
To undo this revert operation:
git revert HEAD~{N}..HEAD
This will create new commits that restore the reverted changes.
Alternatively, if not yet pushed:
git reset --soft HEAD~{N}
git checkout -- .
(Use with caution - this discards the revert commits)| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-22 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-19 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-05 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-07 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-08 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-03 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-02 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-20 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-15 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-09 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-21 | pass→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-18 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-12 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-17 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-16 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-10 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-06 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-01 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-04 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-11 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-13 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-14 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
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 19 counted toward the lift figure. The other 3 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 +32 percentage points is the difference between those two pass rates over the 19 comparable cases.
The per-case answers from this run were removed by the retention sweep, so the case table below shows the verdicts without the text either arm produced. The counts above were recorded at the time and are unaffected. Answers are now kept for 180 days.
Other measured skills in the registry, with their headline benchmark lift.