The Agent Skills Directory

[PROMPT_INJECTION]: Potential surface for indirect prompt injection via terminal configuration files.
Ingestion points: The skill directs the agent to read and modify configuration files located at $XDG_CONFIG_HOME/ghostty/config or ~/Library/Application Support/com.mitchellh.ghostty/config.
Boundary markers: Absent. The instructions do not specify how the agent should distinguish between configuration directives and malicious instructions embedded in comments or values.
Capability inventory: The skill utilizes file system tools such as Read, Edit, Write, Grep, and Glob to manage configuration state.
Sanitization: Absent. There is no mention of validating or escaping configuration content to prevent unintended instruction execution.
[EXTERNAL_DOWNLOADS]: References a configuration schema from the author's GitHub repository.
The documentation points to a JSON schema for syntax validation located at https://raw.githubusercontent.com/sammcj/vscode-ghostty-config-syntax/refs/heads/main/schema/ghostty-config-syntax.schema.json.
[COMMAND_EXECUTION]: Documents administrative CLI commands for the Ghostty terminal.
Provides guidance on using commands such as ghostty +validate-config, ghostty +show-config, and ghostty +edit-config for terminal management.

ghostty-config