rivet-actors
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill architecture is designed to ingest and process data from external clients and message queues. Documentation examples, such as the workflow patterns, demonstrate handling parameters like URLs from these sources which are then used in network operations. While standard for this type of framework, it represents an architectural surface for indirect injection if developers do not implement robust validation.
- Ingestion points: Actor action parameters and message queue bodies (e.g.,
examples/docs/actors-crash-course/workflows.ts). - Boundary markers: Absent in simplified examples, though the framework supports and documents schema-based validation (e.g.,
examples/docs/actors-actions/schema-validation.ts). - Capability inventory: The framework supports network requests (
fetch), persistent state management (KV), and database operations (SQLite). - Sanitization: Examples demonstrate using libraries like
zodfor input validation and manual parameter cleaning. - [SAFE]: No malicious patterns, persistence mechanisms, or unauthorized privilege escalations were detected. All network operations and dependencies point to official vendor resources or established developer tools.
Audit Metadata