The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides numerous shell commands for interacting with WordPress installations via WP-CLI, including scripts that execute arbitrary PHP code for database operations.
Evidence: references/wp-cli-operations.md contains commands such as wp eval " \$data = file_get_contents('/tmp/elementor-content.json'); ... update_post_meta(<post_id>, '_elementor_data', wp_slash(\$data)); ".
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and manipulate Elementor JSON data, which may contain untrusted content like HTML widgets, custom CSS, or shortcodes from external sources.
Ingestion points: Data is imported from JSON files using wp elementor library import or read from the database using wp post meta get.
Boundary markers: The workflow includes instructions to parse payloads, recursively check required fields, and confirm root types before performing write operations.
Capability inventory: The agent has the ability to execute shell commands, run PHP code via wp eval, and perform file system operations for backups.
Sanitization: SKILL.md and references/best-practices.md explicitly instruct the agent to treat rich text, URLs, custom attributes, and shortcodes as untrusted input and to avoid inventing executable content.

elementor-content