cloudkit

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill follows established Apple developer patterns for data synchronization and conflict resolution. It correctly uses secure coding practices for data persistence (NSSecureCoding) and error handling (CKError). All code examples utilize standard system frameworks without external or unverifiable dependencies.
  • [INDIRECT_PROMPT_INJECTION]: The skill describes processing data from external sources (CloudKit databases), which inherently involves untrusted data ingestion. However, the risk is mitigated as the instructions focus purely on data synchronization and merge logic rather than using the content for instruction execution.
  • Ingestion points: Records are fetched via CKDatabase.records(matching:) in SKILL.md and tracked via CKFetchRecordZoneChangesOperation in references/cloudkit-patterns.md.
  • Boundary markers: None present in code snippets.
  • Capability inventory: File system operations via FileManager.setUbiquitous and CKAsset management; UI presentation via UICloudSharingController.
  • Sanitization: None present; the skill prioritizes three-way merge logic for data integrity over content validation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 03:42 PM
Security Audit — agent-trust-hub — cloudkit