The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill provides documentation and examples for using the Bricks Builder 'code' element, which allows for the execution of custom CSS and JavaScript at runtime within the WordPress environment.
Evidence found in 'bricks-layouts.md' and the 'bricks-layouts/' subfolder, where element nodes include 'javascriptCode', 'cssCode', and 'executeCode: true' settings.
The documentation in 'bricks-builder-permissions.md' also discusses user capability requirements for these features, which serves as a mitigation.
[INDIRECT_PROMPT_INJECTION]: The skill defines a significant attack surface for indirect prompt injection by documenting how the agent should ingest and process external data via dynamic tags and form submissions.
Ingestion points: Untrusted data enters the agent context through form field definitions in 'bricks-forms.md' and through dynamic data provider tags (e.g., '{acf_*}', '{post_content}') in 'bricks-dynamic-data.md'.
Boundary markers: There are no explicit boundary markers or 'ignore embedded instructions' warnings provided in the JSON patterns to prevent data from being interpreted as instructions by the builder or subsequent processing steps.
Capability inventory: The skill documents capabilities to execute JavaScript and perform network operations via the 'webhook' form action and 'code' elements.
Sanitization: While the documentation mentions that the Bricks platform performs sanitization and code signing, the instructions for creating custom dynamic data providers ('bricks-dynamic-data.md') allow for raw data output which could bypass protections if implemented incorrectly.

bricks