The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell scripts to detect available model CLI tools and execute them. These operations are part of the skill's primary function to orchestrate multiple AI models. The scripts utilize temporary directories for isolation and implement basic error handling for missing dependencies.
[EXTERNAL_DOWNLOADS]: The skill mentions external dependencies for optional panelists (codex and gemini CLIs). The detection script checks for their presence on the system before attempting to use them, which is a standard discovery mechanism.
[REMOTE_CODE_EXECUTION]: While the skill executes shell commands to run AI panelists, these commands are targeted at local CLI tools provided by reputable model providers. The execution is scoped to temporary directories to prevent unintended file modifications.
[PROMPT_INJECTION]: The skill instructions specify passing the user task verbatim to panelists without modification or safety bypasses. It includes explicit rules for model isolation and synthesis, which are functional rather than deceptive.

fusion