Moderation

Intersect Required

This feature requires a cloud or self-hosted Intersect instance.

Content moderation filters harmful, inappropriate, or policy-violating content in AI-powered workflows. Essential for user-facing applications to prevent abuse, toxic outputs, or compliance violations.

How Habits Implements It

Habits uses Intersect Moderation to scan inputs and outputs for harmful content.

Requires: Intersect API key

Enable Moderation

Set up your Intersect credentials in .env:

INTERSECT_URL=https://your-tenant.intersect.site
INTERSECT_API_KEY=your_api_key

Then configure your workflow: { "id": "moderated-workflow", "habits": { "moderation": { "enabled": true, } } }

### Stack Configuration

```yaml
# config: stack.yaml (Cortex configuration, not a habit)
security:
  moderation:
    enabled: true
    provider: intersect
    categories:
      - hate
      - violence
      - sexual
      - self-harm
    action: block  # or 'flag'
    threshold: 0.7

Configuration Options

Option	Values	Description
action	block, flag	Block request or flag for review
threshold	0.0 - 1.0	Sensitivity level
categories	array	Content categories to check

Flagged or blocked content is logged for audit purposes.