fix(guardrail): report disabled video content-safety stage as skipped - #172
Merged
lfengad merged 3 commits intoAug 12, 2026
Merged
Conversation
The video content-safety classifier is intentionally disabled this release (commented out in presets.py), but the runtime log line 'No safety models found, returning safe' reads as a successful check that passed, so operators who enable guardrails believe video screening is active. Clarify the warning so a skipped check is reported as skipped: 'No content-safety models configured: the safety check is DISABLED...' applies to any runner with no models configured (text and video).
lfengad
approved these changes
Aug 10, 2026
Dinghow
approved these changes
Aug 12, 2026
lfengad
approved these changes
Aug 12, 2026
lfengad
enabled auto-merge (squash)
August 12, 2026 08:51
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Addresses #170, low-effort option from the issue: make the disabled video content-safety stage unmistakable.
GuardrailRunner.run_safety_checkalready warns when no safety models are configured, but the message ("No safety models found, returning safe") reads as a successful check that passed, which is exactly the silent-posture gap reported: withVideoContentSafetyFiltercommented out, guardrails-enabled generation logs a warning that looks like a passing verdict.Now (
cosmos_framework/auxiliary/guardrail/common/core.py):Behavior is unchanged — only the diagnostic message is precise now, so operators can see the stage is skipped.