constructive-entities

Pass

Audited by Gen Agent Trust Hub on Aug 27, 2026

Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [DYNAMIC_EXECUTION]: The framework generates database schema objects, RLS policies, and triggers at runtime based on structured JSON blueprint definitions provided via CLI or ORM.
  • [INDIRECT_PROMPT_INJECTION]: The provisioning logic ingests complex blueprint JSONB data, which serves as an attack surface for indirect prompt injection if the source definitions are untrusted.
  • Ingestion points: Blueprint definitions processed during Phase 0 as described in references/blueprint-entity-types.md.
  • Boundary markers: The system implements defined execution phases (Phase 0, 1, and 2) to scope infrastructure generation.
  • Capability inventory: Support for creating database tables, storage buckets, and security policies via the constructive CLI and ORM mutations.
  • Sanitization: The documentation references the tg_validate_blueprint_definition database trigger used to validate blueprint schemas during ingestion.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 27, 2026, 01:07 PM
Security Audit — agent-trust-hub — constructive-entities