dineway-building-site

Pass

Audited by Gen Agent Trust Hub on Sep 12, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the creation of a CMS-driven site that ingests content from external sources (JSON seed files and database collections) and renders it into pages. This architecture presents an attack surface for indirect prompt injection if malicious instructions are embedded in the site content.
  • Ingestion points: Ingests schema and content from seed/seed.json and database entries retrieved via getDinewayCollection and getDinewayEntry functions.
  • Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded content within the data templates.
  • Capability inventory: The skill uses the npx dineway CLI for management tasks and performs network downloads for media assets.
  • Sanitization: The skill provides a reference example for an RSS feed that includes an escapeXml function to sanitize string data before output.
  • [EXTERNAL_DOWNLOADS]: The skill fetches a JSON schema from https://dineway.foodism.ai/seed.schema.json for validation and uses a $media property in seed files to download images from remote URLs. These are functional downloads necessary for the CMS's operation and population.
  • [COMMAND_EXECUTION]: Utilizes the npx dineway CLI for various development and administrative tasks, such as npx dineway dev, npx dineway seed --validate, and npx dineway deploy forgeway. These commands are specific to the framework's tooling and deployment workflow.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 12, 2026, 02:47 PM