The Agent Skills Directory

[COMMAND_EXECUTION] (LOW): The Rust backend uses std::process::Command to execute a local binary named 'opencode'. While the implementation uses the secure .args() method to prevent shell injection, the reliance on a non-standard external tool increases the risk surface.
[EXTERNAL_DOWNLOADS] (LOW): An automated scanner detected a blacklisted URL associated with the 'main.rs' file referenced in the project structure. Although the file's content is not provided for verification, the presence of a security alert on a core component is a notable risk factor.
[PROMPT_INJECTION] (LOW): The skill is susceptible to indirect prompt injection via the 'prompt' parameter in the 'spawn_opencode' command. Evidence Chain: 1. Ingestion point: 'prompt' argument in spawn_opencode function (lib.rs). 2. Boundary markers: Absent. 3. Capability inventory: Subprocess execution via std::process::Command. 4. Sanitization: Absent.
[DATA_EXFILTRATION] (SAFE): The 'get_opencode_db_path' function returns the path to a local database in the user's home directory (~/.opencode/opencode.db). Exposing this path is a functional requirement for the application's local state management and does not constitute unauthorized exfiltration.

tauri-solidjs