Installs into .claude/skills of the current project.
Are you the author of Api Platform Serialization?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/dev-toolings-api-platform-serialization)
---
name: api-platform-serialization
description: "Control API Platform serialization with groups, #[Context], IRI links (readableLink/writableLink), and custom context builders"
capabilities: [read, search, edit, shell]
tags: [api-platform]
# projected by `bun run build` — do not edit by hand
allowed-tools:
- Read
- Glob
- Grep
- Write
- Edit
- Bash
---
# Api Platform Serialization (Symfony)
## Use when
- Operations must expose different fields (list, read, create, update).
- A related resource must render as an IRI or be embedded.
- A response needs computed or role-dependent fields.
- A property needs normalizer options such as a date format.
## Default workflow
1. Name groups `entity:operation` and set `normalizationContext` and `denormalizationContext` per operation.
2. Tag properties with `#[Groups]` from the Serializer `Attribute` namespace, not the old `Annotation` one.
3. Control relations with `#[ApiProperty(readableLink: false, writableLink: false)]` to force IRIs.
4. Apply per-property options with `#[Context]` instead of writing a normalizer.
5. Add computed fields in a custom normalizer, and dynamic groups in a decorator of `api_platform.serializer.context_builder`.
6. Prevent cycles with `#[MaxDepth]` and `enable_max_depth`, and hide secrets with `#[Ignore]`.
## Guardrails
- Use separate read and write groups: for example `password` only in the create group, `id` only in read groups.
- A related resource is embedded only when it shares a group with the parent; otherwise it renders as an IRI.
- `#[MaxDepth]` has no effect unless `enable_max_depth` is true in the context.
- Keep entities clean: computed fields belong in normalizers.
## Progressive disclosure
- Use this file for execution posture and risk controls.
- Open references when deep implementation details are needed.
## Output contract
- The groups and contexts per operation, as a field matrix.
- Normalizers or context builders added, and how they are registered.
- Tests or OpenAPI output showing the resulting fields.
## References
- `reference.md`