install
Warn
Audited by Gen Agent Trust Hub on Nov 7, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPRIVILEGE_ESCALATIONPROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill runs
make allon source code located in the plugin directory. This involves executing arbitrary commands defined in a Makefile without validation of its contents. - [PRIVILEGE_ESCALATION]: The script attempts to copy files into the
/Applications/directory, which is a system-protected path on macOS. This typically requires elevated permissions. - [PROMPT_INJECTION]: The agent is instructed to "document the bypass" if macOS Gatekeeper or other permission systems interfere with the installation, encouraging the circumvention of system security features.
- [DYNAMIC_EXECUTION]: The skill identifies as "Self-Evolving" and directs the agent to rewrite the
SKILL.mdfile itself to address failures. This creates a mechanism for the agent to modify its own instructions based on the outcomes of external commands.
Audit Metadata