The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted WordPress content from SQL dumps or REST API responses. This represents a potential surface for indirect injection attacks where malicious instructions could be embedded in post content.
Ingestion points: Data enters via scripts/wp-inventory.mjs and scripts/wp-export.mjs using the loadDump and RestSource classes.
Boundary markers: Rich text content is delimited and cleaned using a custom sanitizeHtml function that employs a strict tag and attribute allowlist.
Capability inventory: The skill has the capability to write JSON files to the local file system, perform network reads via fetch, and execute local Node.js scripts.
Sanitization: Extensive filtering logic in scripts/lib/wp-core.mjs (e.g., SECRET_OPTION, SECRET_FIELD_NAME) and scripts/lib/php-unserialize.mjs (repairing broken serialized strings) ensures sensitive data is removed before processing.
[COMMAND_EXECUTION]: The SKILL.md file provides instructions and specific command-line examples for the agent to execute the migration scripts. These commands are restricted to the local execution of the provided Node.js tools and are intended for site analysis and data export tasks within the migration workflow.
[DATA_EXFILTRATION]: The skill performs network operations to fetch data from target WordPress sites via their REST API. This is the primary purpose of the skill and is protected by a redactUrl utility in scripts/lib/content-common.mjs which prevents sensitive parameters (like tokens or keys) from being stored in the cache or printed in error reports.

migrate-wordpress