All guides
No guide matches your search.
Gutenberg blocks
Find why a block shows an invalid-content error, loses its data, renders wrongly or breaks in the editor.
On this page
When to use it
Use this skill to build, debug or review Gutenberg blocks. It covers block.json registration, invalid-content errors and deprecations, dynamic render.php, InnerBlocks, the Interactivity API, Block Bindings and the iframe editor.
It does not cover theme.json, templates or general plugin code.
Related skills
Other topics go to a neighbor skill.
- Theme templates,
theme.jsonand pattern files:wp-devkit-theme-development - Admin screens and settings pages:
wp-devkit-admin-ui-development - REST routes:
wp-devkit-rest-api-development - Plugin lifecycle:
wp-devkit-plugin-development - Escaping or authorization questions beyond the block renderer:
wp-devkit-security-review - Editor and visual polish:
wp-devkit-ui-design
Run a review with Claude
With the plugin loaded, start a read-only review of a bounded target:
/wp-devkit:block-review <target-path>The short form /wp-devkit:block requests a quick scan.
Run a review with Codex, ChatGPT or Antigravity
Name the installed skill and keep the request small. Replace the target with your files or journey.
$wp-devkit-block-development
Review the target without editing files.
Read the project instructions and the skill's engineering contract first.
Separate confirmed findings from candidates and state the evidence and checks you could not run.The engineering contract is the set of rules every DevKit skill follows, such as staying read-only during a review and sorting findings by how well they are proven.
For the best result, give the block name, the oldest and newest WordPress version you support, the block.json, the build command, and one saved-content sample for each older form of the block. Say which action fails: insert, save, reopen, front end or a second instance.
Reference files the skill loads
The skill reads only the files that match your problem.
block-development-workbook.mdholds the decision table, the symptom table, common false alarms, severity guidance and acceptance checks.metadata-registration-and-build.mdcovers block metadata, registration on the server and in the editor, and build output.saved-content-and-deprecations.mdcovers saved markup, validation and deprecations for older content.dynamic-rendering-and-security.mdcovers server-side rendering, escaping and caching.editor-and-iframe-compat.mdcovers editor behavior, data loading and the iframe editor.interactivity-api.mdcovers interactive blocks.bindings-and-php-only-blocks.mdcovers Block Bindings and blocks written only in PHP.
How a review runs
- The skill follows the metadata from source to build output to server and client registration. It checks asset field types,
file:paths,*.asset.phpand the apiVersion against the range you support. - It traces each attribute from its control to its stored form (comment JSON or markup
source) and on to the render. It decides between a static and a dynamic block from the need for fresh output and the storage cost. - It compares each historical sample with the current
savefunction and attribute definitions, including nested blocks. - For dynamic output, it traces attributes and context to the place where they are escaped. It confirms the wrapper attributes and the empty states.
- It checks editor behavior under the iframe, data loading, effects and errors.
- For interactive blocks, it checks state against context, server initialization and event handling. For bindings or PHP-only blocks, it reads the matching reference.
- It reports the boundary that fails, with the old-content and new-content checks you need. It does not rebuild assets or rewrite saved content during a review.
What you get back
The report leads with the result and the scope reviewed. Each confirmed concern lists the file and line, who triggers it, the reachable path, the impact, the confidence, the smallest fix and a regression check. Candidates and missing evidence are listed apart. The skill writes "not established" when a sample, the build output or a version is missing.
An implementation report lists the changed boundaries, the deprecations added, the commands with exit codes, the samples used, the checks not run and the remaining risk. A passing build or editor test does not prove that old content is safe, that the front end renders correctly, or that every WordPress version works.
Note
Reviews are read-only. A clean review covers the inspected scope and is not a certification. Ask for implementation as a separate, bounded task with a regression check.
Where to go next
Use review to fix for a bounded implementation and Doctor and reports for environment and evidence tools.
Was this page helpful?