salla-app-functions
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill describes a workflow for creating handlers that process external store events (e.g.,
order.created). These events represent an ingestion point for untrusted data that could contain malicious instructions targeting the AI agent or the handler logic. - Ingestion points: Handlers ingest
payload.datafrom Salla store events (mentioned in Step 1). - Boundary markers: Not explicitly defined in this router skill; implementation details are delegated to specialized step-skills like
salla-app-functions-handler. - Capability inventory: The skill utilizes the
salla_functionstool tosave(deploy) andpreview(execute) code in a sandboxed environment. - Sanitization: Not addressed in this high-level router; the workflow includes a dedicated validation step (
salla-app-functions-validate). - [SAFE]: All external references and documentation links point to official vendor domains (
docs.salla.dev,portal.salla.partners) or verified community channels (t.me/salladev). - [SAFE]: The instructions prioritize security-conscious design patterns, such as the use of a sandboxed V8 runtime and pre-authenticated API calls to eliminate the need for manual credential handling (e.g.,
Authorizationheaders).
Audit Metadata