Skills
Navigate
skills/electric-sql/electric/entity-stream-queries/Gen Agent Trust Hub

entity-stream-queries

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill documents methods to read from collections such as inbox, which ingest untrusted data from external sources. While this establishes a potential surface for indirect prompt injection, the provided examples focus purely on database query logic. 1. Ingestion points: Internal database collections manifests and inbox (referenced in references/collections.md). 2. Boundary markers: None mentioned in these technical reference files. 3. Capability inventory: Standard agent runtime commands ctx.send and ctx.spawn (referenced in references/common-queries.md). 4. Sanitization: Not applicable to the database query logic described.
  • [SAFE]: No malicious patterns, unauthorized command execution, or data exfiltration attempts were detected. The skill instructions are limited to technical documentation for vendor-owned state management libraries.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 08:49 PM