paperkit
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEMETADATA_POISONINGINDIRECT_PROMPT_INJECTION
Full Analysis
- [METADATA_POISONING]: The skill metadata and technical documentation claim support for 'iOS 26' and 'Swift 6.3'. Since these versions are currently non-existent, this represents misleading metadata that could cause an agent to assume the presence of system capabilities that are unavailable or generate non-functional code for the current environment.
- [INDIRECT_PROMPT_INJECTION]: The skill provides implementation patterns for a document format (PaperMarkup) that extracts text content from untrusted data sources without proper boundaries or sanitization.
- Ingestion points: The skill demonstrates loading PaperMarkup from external binary data in 'references/paperkit-patterns.md'.
- Boundary markers: The provided code snippets do not include delimiters or instructions to ignore potential commands embedded in the markup text, such as those extracted via the 'indexableContent' property.
- Capability inventory: The skill patterns include file system access for saving data and the ability to render content to images, providing a potential impact path for successful injections.
- Sanitization: No sanitization or validation is performed on the extracted text data before it is presented to or processed by the agent.
Audit Metadata