replatform
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill generates and runs local Node.js and Bash scripts (e.g.,
run-import.js,mint-token.sh) under the project'smigrations/directory to facilitate data extraction and import tasks. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted business data from external source platforms which is then interpreted by the LLM to make semantic mapping decisions.
- Ingestion points: Data is collected from source URLs or CSV files using adapter scripts like
wp-discovery.jsandcsv-discovery.js. - Boundary markers: The workflow does not explicitly mention the use of delimiters or 'ignore embedded instructions' warnings in the prompts generated for the mapping phase.
- Capability inventory: The skill maintains capabilities for shell execution, file system writes, and network operations to perform the migration.
- Sanitization: Includes a
SAFE_MODEfeature that automatically replaces outbound emails and phone numbers with mock data to prevent accidental communication during the migration process. - [EXTERNAL_DOWNLOADS]: Scaffolding and authentication tasks involve downloading official Wix CLI tools and project templates from the NPM registry at runtime.
- [REMOTE_CODE_EXECUTION]: The skill executes vendor-provided logic via
npxandnpm createcommands during the project setup phase to scaffold headless sites.
Audit Metadata