What a skill folder adds to a command file
Files in .claude/commands/ keep working and accept the same frontmatter as a skill, except name and paths. A skill adds a directory for supporting files, frontmatter that decides whether you or Claude may invoke it, and automatic loading when a request matches its description. When both share a name the skill runs, and the documentation recommends skills for new work.
Typing the name, or letting Claude choose
Both routes are open by default. Typing /skill-name at the start of a message runs the skill, and whatever follows the name reaches it as $ARGUMENTS. disable-model-invocation: true makes it manual only, which the documentation suggests for side effects such as /deploy or /commit, and keeps its description out of context. user-invocable: false does the opposite: only Claude can call it.
When a plugin is worth the extra layer
Skills, subagents, hooks and MCP servers all work without a plugin. A plugin ships several of them as one unit, installed with one command and updated from its marketplace. Its skills carry the plugin name as a prefix, so they never displace one of yours. While it is enabled, each skill Claude can invoke on its own adds its name and description to context on every turn, even in sessions that never use it.