skills/wix/skills/wix-replatform/Gen Agent Trust Hub

wix-replatform

Pass

Audited by Gen Agent Trust Hub on Aug 27, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests complex business data from external REST APIs (WordPress, Shopify) and user-supplied CSV files, representing a potential surface for indirect prompt injection. The skill mitigates this through a structured and validated pipeline.
  • Ingestion points: Data is imported via scripts like wp-discovery.js, csv-discovery.js, and quick-mode.js.
  • Boundary markers: The skill employs NDJSON for record streaming, utilizes machine-readable orchestration JSON for state management, and maintains a 'faithfulness ledger' to track and surface data fidelity issues to the user.
  • Capability inventory: Capabilities include file system manipulation for artifact creation, network access for fetching source data and writing to Wix APIs, and local shell execution for management tasks.
  • Sanitization: Uses specialized build libraries (wix-build.js) for slug sanitization and sanitizeWixRequestBody to replace real contact data with mock values when safe mode is active.
  • [COMMAND_EXECUTION]: The skill performs command execution using standard CLI tools such as @wix/cli for site authentication and npm create @wix/new for project scaffolding. It also provides instructions for using ngrok to manage media reachability during local development. These tools are vendor-owned or well-known utilities utilized for their primary intended functions.
  • [EXTERNAL_DOWNLOADS]: The migration process involves downloading and executing official packages from Wix, including the Wix CLI and Astro-based storefront templates. These downloads are from trusted vendor repositories and are necessary for the skill's primary function of site deployment and data migration.
  • [DYNAMIC_EXECUTION]: The skill includes an environment recovery mechanism in its 1-click mode that attempts to resolve tool paths (like node or yarn) by executing commands within the operator's interactive login shell. This is a targeted implementation to handle shell environment discrepancies and is restricted by guardrails that prevent the use of hardcoded machine-specific paths.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 27, 2026, 03:37 PM
Security Audit — agent-trust-hub — wix-replatform