The Agent Skills Directory

[SAFE]: The skill architecture is built on deterministic and secure workflows for handling sensitive migration data.
Credential Management: Sensitive configuration, such as API keys and site identifiers, is managed through project-local .env files. The skill provides a utility script (mint-token.sh) that acquires short-lived access tokens via the official Wix CLI and writes them directly to these files, ensuring secrets never appear in the agent's conversation transcript.
PII Protection: A robust 'Safe Mode' feature is integrated into the core write runtime (wix-writers.js). When enabled, it identifies email and phone fields in outbound payloads and replaces them with deterministic mock values (e.g., replatform+entity_id@wix.com), preventing accidental contact with real users during the migration process.
Telemetry Privacy: The telemetry recorder (telemetry-recorder.js) implements a multi-pattern mechanical scrubber that automatically identifies and masks potential secrets, JWT tokens, URLs, and email addresses in free-text fields before they are logged or transmitted.
Trusted Supply Chain: Remote operations are limited to well-known vendor resources. The skill uses @wix/cli and @wix/new via npx/npm for site management and interacts with verified endpoints like wixapis.com and frog.wix.com (Wix's event sink), which are identified as safe in the developer context.
User Oversight: The workflow enforces multiple mandatory approval gates, such as the 'Execution plan report' and 'Mapping review summary,' ensuring that the user has full visibility and control over what data is being moved and what changes will be made to the target site before any write operations occur.

replatform