db-core/collection-setup
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of data from external sources (e.g., REST APIs, ElectricSQL, PowerSync) into the agent's context, which constitutes an attack surface for indirect prompt injection.
- Ingestion points: Data is ingested via the
queryFninqueryCollectionOptions(SKILL.md) and shape streams inelectricCollectionOptions(references/electric-adapter.md). - Boundary markers: No explicit natural language boundary markers or instructions to ignore embedded commands are provided for processed data within the prompt templates.
- Capability inventory: The skill possesses network fetching capabilities (
fetch) and database/file system write capabilities across its various adapters (RxDB, PowerSync, LocalStorage) as detailed inSKILL.mdand the adapter reference files. - Sanitization: The skill strongly encourages and provides examples for strict schema validation using libraries like Zod, which serves as a significant mitigation against malformed data and schema confusion (
references/schema-patterns.md).
Audit Metadata