bootstrap-astro-ses
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill implements a contact form that processes untrusted user input and interpolates it into email templates without sanitization, creating a potential surface for indirect content injection.\n
- Ingestion points: The POST handler in src/pages/api/contact.ts ingests untrusted JSON data from the request body.\n
- Boundary markers: While a Zod schema is used in src/shared/schema.ts for type validation, it does not include instructions to ignore embedded commands or provide sufficient delimiters.\n
- Capability inventory: The skill uses the @aws-sdk/client-ses package to send emails, which constitutes a network capability.\n
- Sanitization: In src/pages/api/contact.ts, the functions formatEmailHtml and formatEmailText interpolate user input directly into strings without escaping HTML or filtering malicious content.
Audit Metadata