kmp-jni-pro
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a collection of engineering guidelines and reference documents. It does not include scripts or binaries that execute on the host system. The instructions promote defensive programming practices such as memory management (ReleaseStringUTFChars), RAII, and symbol isolation (RTLD_LOCAL).
- [EXTERNAL_DOWNLOADS]: The skill references an external tool for JNI binding generation at 'github.com/ronjunevaldoz/jni-binding-generator'. This tool belongs to the identified author and is presented as an optional resource for reducing boilerplate. The documentation also points to official Android NDK guides.
- [PROMPT_INJECTION]: The skill uses 'HARD STOP' and 'MANDATORY' markers to enforce its internal development workflow (e.g., Phase 0 Discovery). These are architectural constraints for the agent's persona and do not attempt to bypass system safety filters or override global security instructions.
- [COMMAND_EXECUTION]: The provided code snippets and CMake examples are templates for the developer to use. They include standard build-time commands (grep, nm, cmake) intended for manual auditing and configuration, not for silent execution by the agent to compromise the system.
Audit Metadata