website-checkout

Pass

Audited by Gen Agent Trust Hub on Sep 22, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill utilizes npx -y solvapay@latest init to configure the environment and install necessary SDK packages. This command downloads and executes a script directly from the vendor's package on the NPM registry.
  • [EXTERNAL_DOWNLOADS]: Downloads various official SolvaPay and Supabase packages (e.g., @solvapay/next, @supabase/supabase-js) from the NPM registry as part of the setup process.
  • [COMMAND_EXECUTION]: The verification loop involves running curl commands to test the authentication and checkout endpoints, ensuring both success and failure paths behave as expected.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing external data such as product references and authentication tokens.
  • Ingestion points: API routes (e.g., /api/create-checkout-session, /api/check-access) ingest user-supplied identifiers and JWT tokens.
  • Boundary markers: While not explicitly using delimiters in the generated code snippets, the skill instructs the agent to implement logic that gates premium views based on authoritative server-side checks.
  • Capability inventory: Includes file system operations (package installation and initialization) and network operations (API calls to SolvaPay and Supabase).
  • Sanitization: The instructions emphasize verifying purchase states against the 'server truth' and keeping secret keys on the backend to prevent client-side exposure.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 22, 2026, 10:14 AM