# Mandatory rules

Global mandatory rules apply to every project, including projects using a custom policy. They set a shared content standard alongside your [category actions, allowances and restrictions](/docs/guides/moderation-policies). Rules that depend on where you serve viewers are covered in [Jurisdiction rules](/docs/guides/jurisdiction-rules).

## Global content rules

These are CalmLens service-wide standards. The official sources below explain related legal definitions and harms; their territorial scope, exceptions and legal tests differ from CalmLens’s global rules.

Every policy includes these global mandatory `block` rules:

| Restriction              | Category | What it blocks                                                                                                                                                                                                                  | Official background source                                                                                                                                                                                                 |
| ------------------------ | -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Sexualised minors**    | Sexual   | Sexual content involving a child or a character presented as a child, including drawings, animation and generated images.                                                                                                       | [CPS: indecent and prohibited images of children](https://www.cps.gov.uk/prosecution-guidance/indecent-and-prohibited-images-children)                                                                                     |
| **Intimate image abuse** | Sexual   | A real person's intimate images shared, offered, sought or threatened without their consent, including hidden-camera images, sextortion and sexual deepfakes.                                                                   | [eSafety: image-based abuse](https://www.esafety.gov.au/key-topics/image-based-abuse)                                                                                                                                      |
| **Extreme pornography**  | Sexual   | Sex with animals or a dead body, or sexual acts that threaten life or cause serious injury, including realistic drawn or generated depictions.                                                                                  | [Home Office: extreme pornography](https://www.gov.uk/government/publications/criminal-investigations-indecent-and-obscene-materials/criminal-investigations-indecent-and-obscene-material-accessible#extreme-pornography) |
| **Violent extremism**    | Violence | Propaganda, recruitment, fundraising or praise for a terrorist or violent extremist group or its attacks, and attackers' own footage or manifestos. News, research and counterspeech that don't promote them follow the policy. | [EU Regulation 2021/784, Articles 1(3) and 2(7)](https://eur-lex.europa.eu/eli/reg/2021/784/oj/eng)                                                                                                                        |

Sources checked on 27 October 2026.

## Precedence and safety holds

They appear as mandatory rules for their category, but they can't be edited or removed. They apply even when the category is turned off, a per-asset action override allows the asset, or an allowance covers the content, and each acts once its policy score reaches 50%. An asset matching **Sexualised minors** is also placed under a [safety hold](/docs/guides/asset-lifecycle) while CalmLens reviews it, so its content stays unavailable and can't be deleted until the review ends.

## Read the decision

A matching global rule records `source: "mandatory-rule"` and its stable `rule` key in `classification.moderation.policy.decision`. The rule identifier identifies the rule that acted; it is separate from the content qualifiers used to assess it. See [Mandatory and regional decisions](/docs/guides/classification-results#mandatory-and-regional-decisions) for the report fields.

A global block applies everywhere. [Jurisdiction rules](/docs/guides/jurisdiction-rules) can add a stricter outcome for particular viewers, but cannot soften the global decision.