Inspect code differences between experiment commits.
npx skills add https://github.com/tokyo-megacorp/autoimprove --skill diff-tokyo-megacorp
Please help me install this Agent Skill. Skill: diff Source: https://github.com/tokyo-megacorp/autoimprove/tree/main/skills/diff Command: npx skills add https://github.com/tokyo-megacorp/autoimprove --skill diff-tokyo-megacorp
This Skill enables users to inspect detailed code-level differences between experiment commits, facilitating validation and understanding of changes that impact performance or behavior.
Use the diff skill to view the changes made in experiment 12 by running /autoimprove diff 12.
High-intent search queries and answers about installing and using this skill.
To view code differences from experiment commits, you can use automated diff filtering and summarization to inspect specific changes without running manual git commands. This allows you to quickly assess code evolution and validate modifications.
Diff summarization across multiple experiments works by aggregating code differences from a range of experiment commits into a single comparison. This provides an overall view of the cumulative impact from a series of modifications.
Yes, you can review code changes for a specific experiment before a rollback by viewing the actual code differences introduced in that commit. This helps verify exactly what code was altered to inform your decision.
The best way to compare code evolution across version control experiments is to use an automated diff tool that handles filtering and comparison. This bypasses manual git commands and provides dynamic views of changes across kept experiments.
No, you do not need to run manual git commands to inspect code regressions. Automated diff viewing handles the filtering and comparison of experiment commits, allowing you to understand regressions through summarized code differences directly.
TimoFreibergOrchestrate development, review, and fix loops with specialized subagents.
arbazkhan971Finalize feature branches via merge, pull request, keep, or discard after build, lint, and tests.
ollieb89Evaluate LLM outputs with automated metrics and LLM-as-Judge workflows.
skuridinGenerate standardized GitHub pull request titles and descriptions from commit history.
