diff --git a/.buildkite/scripts/pipelines/pull_request/pipeline.ts b/.buildkite/scripts/pipelines/pull_request/pipeline.ts
index 4c071df810e5c..6ab844204c448 100644
--- a/.buildkite/scripts/pipelines/pull_request/pipeline.ts
+++ b/.buildkite/scripts/pipelines/pull_request/pipeline.ts
@@ -644,12 +644,13 @@ const SKIPPABLE_PR_MATCHERS = prConfig.skip_ci_on_only_changed!.map((r) => new R
 
     // Run Saved Objects checks conditionally
     if (
-      await doAnyChangesMatch([
+      !GITHUB_PR_LABELS.includes('ci:skip-so-check') &&
+      (await doAnyChangesMatch([
         /^packages\/kbn-check-saved-objects-cli\/current_fields.json/,
         /^packages\/kbn-check-saved-objects-cli\/current_mappings.json/,
         /^src\/core\/server\/integration_tests\/ci_checks\/saved_objects\/check_registered_types.test.ts/,
         /^\.buildkite\/pipelines\/pull_request\/check_saved_objects\.yml/,
-      ])
+      ]))
     ) {
       pipeline.push(
         getPipeline('.buildkite/pipelines/pull_request/check_saved_objects.yml', cancelable)
diff --git a/packages/kbn-check-saved-objects-cli/README.md b/packages/kbn-check-saved-objects-cli/README.md
index 620b15241060b..66603b87689aa 100644
--- a/packages/kbn-check-saved-objects-cli/README.md
+++ b/packages/kbn-check-saved-objects-cli/README.md
@@ -17,6 +17,10 @@ On a PR, the step:
 1. **Resolves a baseline commit**: It uses the PR merge-base commit (`GITHUB_PR_MERGE_BASE`). Because a snapshot may not exist for that exact SHA, the script walks the commit’s parent chain (up to a limit) until it finds a commit for which a snapshot exists in the cloud bucket (`findExistingSnapshotSha`).
 2. **Runs the check**: It calls `node scripts/check_saved_objects --baseline <resolved-sha> --algorithm both` (and `--fix` when auto-commit is enabled). The script fetches the baseline snapshot from the bucket, starts Kibana to build the current type registry, and runs validations comparing current state to that baseline.
 
+### Skipping the check
+
+Add the `ci:skip-so-check` label to a PR to bypass the "Check changes in Saved Objects" step entirely. Use this as an escape hatch for unrelated false positives (for example, `Joi` version bumps or no-op schema refactors that change snapshot hashes without changing the persisted shape). Do not use it to bypass real Saved Object type changes.
+
 ### Where snapshots come from
 
 Baseline snapshots are **stored in a GCP bucket** and served at: