The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process potentially untrusted data from local configuration files and external theme repositories.
Ingestion points: Reads user-defined configuration files in ~/.config/ (e.g., Hyprland, Waybar) and processes output from command discovery tools.
Capability inventory: The skill has the capability to execute administrative commands via the omarchy CLI, systemctl, and pacman, and can write to various filesystem paths.
Boundary markers: There are no explicit instructions to use delimiters or specific prompt engineering techniques to protect the agent from instructions embedded within the configuration files it reads.
Sanitization: The instructions do not define methods for sanitizing or validating the content of the files or repository metadata before it is processed.
[COMMAND_EXECUTION]: The skill frequently interacts with the operating system through high-privilege command-line tools.
Evidence: The documentation describes workflows for managing packages via pacman, services via systemctl, and desktop state via hyprctl and omarchy wrappers.
Mitigation: The skill mitigates risks by requiring interactive terminals for sudo operations and mandatory user confirmation for destructive or system-wide actions like reinstallation or snapshots.
[DATA_EXFILTRATION]: The diagnostic features of the skill involve gathering and potentially transmitting sensitive system information.
Evidence: The documentation for omarchy debug acknowledges that the command generates logs containing hardware and journal information and may offer an upload action.
Mitigation: The instructions explicitly forbid the agent from uploading or sharing logs without express user authorization and advise local inspection and redaction of sensitive details.

omarchy-help