The Agent Skills Directory

[COMMAND_EXECUTION]: The skill generates and instructs the agent to execute potentially dangerous shell commands, including rm -rf for directory removal and docker volume rm for resource deletion.
[PRIVILEGE_ESCALATION]: The skill explicitly includes instructions for system-level cleanup tasks that require sudo privileges, such as clearing /Library/Caches and /var/log. While it advises the user to run these manually, it provides the full command strings for execution.
[EXTERNAL_DOWNLOADS]: The skill directs the agent to install an external CLI tool, mole, from a third-party Homebrew tap (tw93/tap/mole) which is not a verified vendor repository.
[INDIRECT_PROMPT_INJECTION]: The skill has a significant attack surface for indirect prompt injection as it processes untrusted data from the local filesystem.
Ingestion points: The skill reads file names, directory paths, and metadata (size, age) from the user's disk via mo analyze and custom scripts like analyze_large_files.py.
Boundary markers: There are no explicit instructions or delimiters mentioned to prevent the agent from interpreting maliciously crafted filenames or file content as instructions.
Capability inventory: The skill possesses file deletion capabilities (rm), privilege escalation patterns (sudo), and interactive shell management via tmux.
Sanitization: No sanitization or validation of the ingested filesystem data is described before it is displayed to the user or interpolated into cleanup commands.

macos-cleaner