vtex-io-session-apps

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill provides legitimate technical documentation and code patterns for VTEX IO session management. The instructions focus on correct platform usage, including namespace ownership, input-vs-output field declarations, and proper caching strategies using LRU and VBase.
  • [INDIRECT_PROMPT_INJECTION]: The skill describes a data ingestion surface where external session data from ctx.request.body is processed by transform handlers.
  • Ingestion points: Untrusted data enters the transform handler via ctx.request.body in the provided transform.ts code snippet.
  • Boundary markers: The skill advocates for strict schema contracts via vtex.session/configuration.json to define allowed inputs and outputs.
  • Capability inventory: The transform logic can perform network operations (fetchCostCenterFromAPI) and write to persistent storage via VTEX VBase (ctx.clients.vbase).
  • Sanitization: The skill mentions an input parsing step (parseSfpInputs) and advocates for using authoritative private namespaces as the read model to prevent reliance on potentially tainted public inputs.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 07:53 PM