core-data

Pass

Audited by Gen Agent Trust Hub on Sep 11, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of data from external storage, creating a potential surface for indirect prompt injection.\n- Ingestion points: Managed object data is retrieved via NSFetchRequest and NSPersistentHistoryChangeRequest as described in SKILL.md, references/core-data-patterns.md, and references/persistent-history.md.\n- Boundary markers: None provided; the skill lacks specific delimiters or warnings to ignore instructions embedded in stored data.\n- Capability inventory: The skill includes database write access (context.save()), batch deletion/insertion (context.execute()), and persistent history management across multiple files as shown in references/core-data-patterns.md and references/persistent-history.md.\n- Sanitization: None provided; there is no guidance on sanitizing or validating data retrieved from the persistent store before it enters the agent's context.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 11, 2026, 11:13 AM
Security Audit — agent-trust-hub — core-data