Tutorial: Content Safety editor¶
CSA Loom
content-safetyeditor — verified working against a live console by the UAT harness on 2026-07-01.
Open the editor¶
- Sign in to your CSA Loom Console (for example
https://<your-console-host>). - Open or create a workspace from the Workspaces page.
- Click + New item and choose Content Safety from the catalog.
- The editor opens at
/items/content-safety/<id>:
What this editor does¶
Content Safety is Azure AI Content Safety — text and image moderation across hate/violence/sexual/self-harm with severity thresholds. In Loom you configure thresholds and wire it in front of any LLM call.
Getting started¶
- Set categories — Enable the harm categories you want screened (hate, violence, sexual, self-harm).
- Tune severity thresholds — Set the severity threshold per category that should block content.
- Test content — Run sample text or images through to see the severity scores.
- Wire in front of the LLM — Place Content Safety ahead of prompt flow or agent calls to filter inputs and outputs.
Learn more¶
- Microsoft Learn reference: https://learn.microsoft.com/azure/ai-services/content-safety/overview
Verified by the UAT harness¶
- Tested at:
2026-05-26T13:54:35.221Z - Verdict: A (renders cleanly, real backend responded)
- Test source:
apps/fiab-console/e2e/editors.uat.ts
