widgetkit
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSMETADATA_POISONINGINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill links to external documentation hosted on "sosumi.ai" (e.g., "https://sosumi.ai/documentation/widgetkit"). This domain is not an official source for Apple developer documentation and is not a recognized trusted vendor.
- [METADATA_POISONING]: The content repeatedly references "iOS 26+", which is not a current or announced version of the operating system. This provides misleading technical context that may result in unreliable code implementation.
- [INDIRECT_PROMPT_INJECTION]: The skill establishes patterns for processing untrusted external data via WidgetKit providers. 1. Ingestion points: Data is ingested through "AppIntentTimelineProvider" parameters and external network requests in "SKILL.md" and "references/widgetkit-advanced.md". 2. Boundary markers: The provided code snippets do not include delimiters or instructions for the agent to ignore potentially malicious instructions embedded in data. 3. Capability inventory: The skill uses capabilities for network communication and persistent local storage ("UserDefaults", "DataStore"). 4. Sanitization: There is no guidance or example code for validating or sanitizing user-provided parameters or external API responses.
Audit Metadata