moai-lang-cpp
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHCOMMAND_EXECUTIONPROMPT_INJECTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION] (HIGH): The skill is authorized to use the
Bashtool, which provides a significant attack surface for arbitrary code execution on the host system. - [PROMPT_INJECTION] (HIGH): The skill is susceptible to Indirect Prompt Injection. Ingestion points: Reads C++ source files and CMake configurations via
ReadandGrep. Boundary markers: None. The agent is not instructed to ignore comments or instructions within data. Capability inventory: Access to theBashtool. Sanitization: None. A malicious codebase could contain instructions in comments designed to exploit the agent's Bash capability. - [REMOTE_CODE_EXECUTION] (HIGH): The documentation recommends cloning and executing a bootstrap script from a remote repository (
vcpkg). While the source is trusted, this pattern of fetching and running remote scripts is a known RCE vector. - [EXTERNAL_DOWNLOADS] (LOW): The skill uses
FetchContentand references standard library providers like GitHub and Conan. These are downgraded to LOW per the [TRUST-SCOPE-RULE] for trusted sources.
Recommendations
- AI detected serious security threats
Audit Metadata