make-app-service
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill involves processing data from external Make platform APIs and AI assistant interfaces (Artifacts, SSE events), which presents a potential surface for indirect prompt injection if the ingested data contains instructions designed to manipulate the agent.
- Ingestion points: Data is ingested from the Make Meta/Data APIs and AI Assistant Gateway as described in
SKILL.mdandreferences/make-data-adapter.md. - Boundary markers: The instructions mandate strict Service-side validation and request/response normalization, providing a structural boundary for ingested data.
- Capability inventory: The skill has the capability to perform network requests, file upload/download operations, and record modifications via backend adapters.
- Sanitization: Route handlers are required to validate all query, body, and path parameters, returning 400 errors for invalid input to prevent processing of malformed data.
- [SAFE]: The skill incorporates multiple security best practices and safety requirements:
- Mandatory redaction of sensitive information such as Authorization headers, cookies, and API keys in logs.
- Prevention of hardcoded secrets by requiring the use of environment-injected variables like
MAKE_APP_KEY. - Implementation of a secure Service download proxy to avoid exposing raw backend file tokens to the client environment.
- Explicit discouragement of executing shell commands or using local development tools (
makecli) within the production runtime environment.
Audit Metadata