The Agent Skills Directory

[SAFE]: The skill and its associated files (icon and agent configuration) contain no evidence of prompt injection, data exfiltration, or unauthorized command execution. The provided Swift code snippets follow Apple's official documentation and best practices.
[INDIRECT_PROMPT_INJECTION]: The skill describes processes for importing and syncing data from external sources.
Ingestion points: Data is processed through ModelContext during imports and via CloudKit synchronization tasks described in SKILL.md.
Boundary markers: None present in the documentation code snippets, as the focus is on framework implementation rather than input validation.
Capability inventory: The skill utilizes file system writing through SQLite-backed storage and network access via CloudKit APIs.
Sanitization: The framework's use of strongly-typed models and the Swift language's memory safety features provide inherent protections against common data-driven attacks.
[METADATA_POISONING]: The skill includes future-dated references (iOS 26+ and 2025-11-28). These appear to be placeholders within the 'Axiom' project context and do not represent a malicious attempt to deceive the agent's security protocols.

axiom-swiftdata