Plugins can choose the effort level for model.fork, and hook matchers match MCP tool names

What

$.model.fork lets a plugin ask the model a side question against the current conversation. It now takes an optional effort setting. Effort controls how much work the model puts into its answer.

• $.model.fork now accepts { prompt, effort? }. Before, it accepted only { prompt }, and the prompt must still be non-empty.
• effort must be one of the known effort levels. Any other value makes the call fail.
• The call now reports which effort level was actually sent. Before, it returned only the usage figures.
• Hook matchers have extra handling for MCP tools. A hook is code that runs at set points, such as before a tool is used, and a matcher decides which tools it applies to. MCP servers are outside programs that give Claude extra tools. A matcher can now match an MCP tool by its full name.

Why

Plugin authors can now pick how much effort a forked question gets, and can see which level was really used. Hook authors can target a specific MCP tool by writing its full name in a matcher.