tauri-architecture-patterns
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides defensive coding examples, such as blocking access to sensitive system directories (e.g.,
/.ssh/,/etc/) and sanitizing filenames in Rust to prevent path traversal or injection attacks. - [SAFE]: Recommendations for credential management are provided, advising against storing API keys in local JSON files and instead using the OS-level keychain via the
keyringcrate. - [INDIRECT_PROMPT_INJECTION]: The skill includes ingestion points where it reads Rust and frontend manifests, configuration files, and application setup from the local repository (SKILL.md). While there are no specific boundary markers to delimit these inputs, the agent's capability inventory includes modifying the repository and designing new features. Sanitization for the agent's own context is absent, though the skill provides extensive sanitization logic for the application it helps build. The risk is considered low due to the educational nature of the skill and the context of manual architectural review.
Audit Metadata