calling-rust-from-tauri-frontend
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill provides code templates for commands like
read_file,open_file, andstream_filethat accept arbitrary file paths directly from the frontend to perform file system read operations (std::fs::read,std::fs::read_to_string,tokio::fs::File::open). Without path validation or sandboxing, these patterns enable arbitrary file read vulnerabilities where the frontend could access sensitive system files. - [DATA_EXFILTRATION]: The
fetch_datacommand example performs network requests viareqwest::getbased on a user-provided URL string, creating a surface for Server-Side Request Forgery (SSRF) where the backend might be induced to make requests to internal or unintended resources. - [INDIRECT_PROMPT_INJECTION]: The skill demonstrates functions that ingest untrusted data from external sources which could influence agent behavior if processed further.
- Ingestion points: Ingests file contents in
open_fileandstream_file, and network responses infetch_data. - Boundary markers: No delimiters or safety warnings are included in the instruction templates.
- Capability inventory: Includes file system read operations and network GET requests.
- Sanitization: No input validation, path normalization, or output sanitization is demonstrated in the provided examples.
Audit Metadata