mobile-storage-sqlite-powersync

Pass

Audited by Gen Agent Trust Hub on Nov 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The documentation and code examples consistently demonstrate the use of parameterized SQL queries (e.g., powersync.execute(sql, [params])). This approach is a standard security best practice that prevents SQL injection attacks by ensuring that user-provided data is never directly concatenated into executable SQL strings.
  • [EXTERNAL_DOWNLOADS]: The skill references well-known, legitimate packages from the PowerSync ecosystem, including @powersync/react-native, @powersync/react, and @powersync/op-sqlite. These dependencies are necessary for the synchronization and database functionality described.
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the synchronization of remote data into a local SQLite environment, creating a potential surface for indirect prompt injection.
  • Ingestion points: Untrusted data enters the agent's context through the PowerSyncBackendConnector and the background synchronization stream, which populates the local SQLite tables.
  • Boundary markers: The skill relies on the structured nature of the SQLite schema rather than explicit natural language delimiters.
  • Capability inventory: The skill possesses database execution capabilities via powersync.execute() and network capabilities via the custom backend connector.
  • Sanitization: The implementation uses parameterized queries for all database interactions, which provides effective sanitization against data-driven execution risks.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 9, 2026, 06:27 PM