Revert specific kept experiments using git revert based on experiment logs.
npx skills add https://github.com/tokyo-megacorp/autoimprove --skill rollback-tokyo-megacorp
Please help me install this Agent Skill. Skill: rollback Source: https://github.com/tokyo-megacorp/autoimprove/tree/main/skills/rollback Command: npx skills add https://github.com/tokyo-megacorp/autoimprove --skill rollback-tokyo-megacorp
The rollback skill allows users to revert specific kept experiments in the codebase, ensuring stable and reliable performance by undoing problematic changes.
Use the rollback skill to list all kept experiments or revert a specific experiment by ID with a natural language command.
High-intent search queries and answers about installing and using this skill.
To revert a git commit from an experiment without losing history, you can use a rollback operation that applies git revert to the specific commit. This preserves accountability while undoing problematic changes to restore baseline code stability.
The best way to undo problematic codebase changes from a kept experiment is to execute a targeted rollback using experiment logs. This systematically reverts the specific commit associated with the experiment to maintain reliable project performance.
Yes, you can list all kept experiments before deciding which commits to revert. The rollback process includes a listing feature that shows all eligible experiments, aiding your review and target identification before performing any undo operations.
You need to perform a rollback on your version control experiment workflow when a kept experiment introduces bugs or regressions into the codebase. Reverting the specific commit restores baseline integrity and ensures stable performance.
Yes, reverting an experiment commit updates the baseline project state directly. The rollback process identifies the target commit from experiment logs and systematically updates the project state to ensure a traceable undo operation.
bakwankawaGenerate human-readable timestamped logs with per-step timing breakdowns.
adiomasRun typecheck, lint, tests, and build in deterministic order.
Kang-chenTrack ML experiments and visualize results across runs using Weights & Biases.
niznik-devCreate custom inspect-ai evaluation tasks with experiment-guided and standalone modes.
