The Agent Skills Directory

[COMMAND_EXECUTION]: The skill frequently executes a local bash script (assets/system-check.sh) to perform platform detection and dependency validation.
[EXTERNAL_DOWNLOADS]: The 'automated' mode is explicitly designed to download and install external software packages (Python, CMake, JUCE, etc.) to the local system.
[REMOTE_CODE_EXECUTION]: Automating the installation of third-party dependencies involves the execution of external installers and scripts, which constitutes remote code execution on the host environment.
[PRIVILEGE_ESCALATION]: The skill's error handling documentation mentions offering 'sudo' as a recovery path for installation failures, indicating it may attempt to acquire root-level permissions.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data which presents an attack surface.
Ingestion points: Reads existing configuration from .claude/system-config.json and processes outputs from the system-check.sh script.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are used when processing the configuration file.
Capability inventory: The skill has access to Bash, Read, Write, and Edit tools, allowing for system-level changes and file modifications.
Sanitization: There is no evidence of sanitization or validation logic for the data read from the configuration file before it is used to influence the setup flow.

system-setup