perf(tools): lower default result sizes to cut response tokens

The default limits let broad queries flood the model context. This
change sets smaller defaults. An explicit parameter still overrides each one.

- file_search: max_results 1000 to 50, context_lines 2 to 0.
- ast_query: max_results 100 to 50.
- ast_query compact output omits the node type unless verbose is set.

On a fixed seven-call suite, response size fell from 198529 to 61960 bytes.

internal/server/server.go

@@ -136,7 +136,7 @@ func (s *Server) registerTools() {
 					mcp.Description("Treat pattern as regex (default: true)"),
 				),
 				mcp.WithNumber("context_lines",
-					mcp.Description("Number of context lines around matches (default: 2)"),
+					mcp.Description("Number of context lines around matches (default: 0)"),
 				),
 				mcp.WithNumber("max_results",
 					mcp.Description("Maximum number of results to return (page size for pagination)"),
@@ -252,7 +252,7 @@ func (s *Server) registerTools() {
 				mcp.WithStringItems(),
 			),
 			mcp.WithNumber("max_results",
-				mcp.Description("Maximum number of results to return (default: 100, page size for pagination)"),
+				mcp.Description("Maximum number of results to return (default: 50, page size for pagination)"),
 			),
 			mcp.WithString("format",
 				mcp.Description("Output format: \"compact\" (default, one line per match), \"verbose\" (full code+captures), \"location\" (file:line only)"),