Locator healing patches, script changes, and baseline updates are never committed automatically. Every AI suggestion lands in your approval queue first.
AI does the detection and suggestion. Humans make the call.
When LocatorHealer finds a better selector, it shows you the old vs. new locator with a confidence score. You approve and it's committed — or reject and it's discarded.
When ScriptReview or ResultAnalyzer detects a script-level fix, the proposed diff is shown inline. Approve to update the file, reject to keep the current version.
When a visual diff exceeds threshold and the change is intentional, you approve the new screenshot as the baseline. Rejected diffs block the build.
AI does the work. You stay in control of what gets committed to your test suite.