diff --git a/checkers/checkers.go b/checkers/checkers.go
index 7ce829d36..0c2ebc00c 100644
--- a/checkers/checkers.go
+++ b/checkers/checkers.go
@@ -2,15 +2,9 @@
 package checkers
 
 import (
-	"fmt"
-	"go/ast"
-	"go/build"
-	"go/token"
 	"os"
 
-	"github.com/go-critic/go-critic/checkers/rulesdata"
 	"github.com/go-critic/go-critic/framework/linter"
-	"github.com/quasilyte/go-ruleguard/ruleguard"
 )
 
 var collection = &linter.CheckerCollection{
@@ -23,93 +17,3 @@ var debug = func() func() bool {
 		return v
 	}
 }()
-
-//go:generate go run ./rules/precompile.go -rules ./rules/rules.go -o ./rulesdata/rulesdata.go
-
-func init() {
-	filename := "rules/rules.go"
-
-	fset := token.NewFileSet()
-	var groups []ruleguard.GoRuleGroup
-
-	var buildContext *build.Context
-
-	ruleguardDebug := os.Getenv("GOCRITIC_RULEGUARD_DEBUG") != ""
-
-	// First we create an Engine to parse all rules.
-	// We need it to get the structured info about our rules
-	// that will be used to generate checkers.
-	// We introduce an extra scope in hope that rootEngine
-	// will be garbage-collected after we don't need it.
-	// LoadedGroups() returns a slice copy and that's all what we need.
-	{
-		rootEngine := ruleguard.NewEngine()
-		rootEngine.InferBuildContext()
-		buildContext = rootEngine.BuildContext
-
-		loadContext := &ruleguard.LoadContext{
-			Fset:         fset,
-			DebugImports: ruleguardDebug,
-			DebugPrint: func(s string) {
-				fmt.Println("debug:", s)
-			},
-		}
-		if err := rootEngine.LoadFromIR(loadContext, filename, rulesdata.PrecompiledRules); err != nil {
-			panic(fmt.Sprintf("load embedded ruleguard rules: %v", err))
-		}
-		groups = rootEngine.LoadedGroups()
-	}
-
-	// For every rules group we create a new checker and a separate engine.
-	// That dedicated ruleguard engine will contain rules only from one group.
-	for i := range groups {
-		g := groups[i]
-		info := &linter.CheckerInfo{
-			Name:    g.Name,
-			Summary: g.DocSummary,
-			Before:  g.DocBefore,
-			After:   g.DocAfter,
-			Note:    g.DocNote,
-			Tags:    g.DocTags,
-
-			EmbeddedRuleguard: true,
-		}
-		collection.AddChecker(info, func(ctx *linter.CheckerContext) (linter.FileWalker, error) {
-			parseContext := &ruleguard.LoadContext{
-				Fset: fset,
-				GroupFilter: func(name string) bool {
-					return name == g.Name
-				},
-				DebugImports: ruleguardDebug,
-				DebugPrint: func(s string) {
-					fmt.Println("debug:", s)
-				},
-			}
-			engine := ruleguard.NewEngine()
-			engine.BuildContext = buildContext
-			err := engine.LoadFromIR(parseContext, filename, rulesdata.PrecompiledRules)
-			if err != nil {
-				return nil, err
-			}
-			c := &embeddedRuleguardChecker{
-				ctx:    ctx,
-				engine: engine,
-			}
-			return c, nil
-		})
-	}
-}
-
-type embeddedRuleguardChecker struct {
-	ctx    *linter.CheckerContext
-	engine *ruleguard.Engine
-}
-
-func (c *embeddedRuleguardChecker) WalkFile(f *ast.File) {
-	runRuleguardEngine(c.ctx, f, c.engine, &ruleguard.RunContext{
-		Pkg:   c.ctx.Pkg,
-		Types: c.ctx.TypesInfo,
-		Sizes: c.ctx.SizesInfo,
-		Fset:  c.ctx.FileSet,
-	})
-}
diff --git a/checkers/embedded_rules.go b/checkers/embedded_rules.go
new file mode 100644
index 000000000..e37fdf14a
--- /dev/null
+++ b/checkers/embedded_rules.go
@@ -0,0 +1,103 @@
+package checkers
+
+import (
+	"fmt"
+	"go/ast"
+	"go/build"
+	"go/token"
+	"os"
+
+	"github.com/go-critic/go-critic/checkers/rulesdata"
+	"github.com/go-critic/go-critic/framework/linter"
+	"github.com/quasilyte/go-ruleguard/ruleguard"
+)
+
+//go:generate go run ./rules/precompile.go -rules ./rules/rules.go -o ./rulesdata/rulesdata.go
+
+func init() {
+	filename := "rules/rules.go"
+
+	fset := token.NewFileSet()
+	var groups []ruleguard.GoRuleGroup
+
+	var buildContext *build.Context
+
+	ruleguardDebug := os.Getenv("GOCRITIC_RULEGUARD_DEBUG") != ""
+
+	// First we create an Engine to parse all rules.
+	// We need it to get the structured info about our rules
+	// that will be used to generate checkers.
+	// We introduce an extra scope in hope that rootEngine
+	// will be garbage-collected after we don't need it.
+	// LoadedGroups() returns a slice copy and that's all what we need.
+	{
+		rootEngine := ruleguard.NewEngine()
+		rootEngine.InferBuildContext()
+		buildContext = rootEngine.BuildContext
+
+		loadContext := &ruleguard.LoadContext{
+			Fset:         fset,
+			DebugImports: ruleguardDebug,
+			DebugPrint: func(s string) {
+				fmt.Println("debug:", s)
+			},
+		}
+		if err := rootEngine.LoadFromIR(loadContext, filename, rulesdata.PrecompiledRules); err != nil {
+			panic(fmt.Sprintf("load embedded ruleguard rules: %v", err))
+		}
+		groups = rootEngine.LoadedGroups()
+	}
+
+	// For every rules group we create a new checker and a separate engine.
+	// That dedicated ruleguard engine will contain rules only from one group.
+	for i := range groups {
+		g := groups[i]
+		info := &linter.CheckerInfo{
+			Name:    g.Name,
+			Summary: g.DocSummary,
+			Before:  g.DocBefore,
+			After:   g.DocAfter,
+			Note:    g.DocNote,
+			Tags:    g.DocTags,
+
+			EmbeddedRuleguard: true,
+		}
+		collection.AddChecker(info, func(ctx *linter.CheckerContext) (linter.FileWalker, error) {
+			parseContext := &ruleguard.LoadContext{
+				Fset: fset,
+				GroupFilter: func(name string) bool {
+					return name == g.Name
+				},
+				DebugImports: ruleguardDebug,
+				DebugPrint: func(s string) {
+					fmt.Println("debug:", s)
+				},
+			}
+			engine := ruleguard.NewEngine()
+			engine.BuildContext = buildContext
+			err := engine.LoadFromIR(parseContext, filename, rulesdata.PrecompiledRules)
+			if err != nil {
+				return nil, err
+			}
+			c := &embeddedRuleguardChecker{
+				ctx:    ctx,
+				engine: engine,
+			}
+			return c, nil
+		})
+	}
+}
+
+type embeddedRuleguardChecker struct {
+	ctx    *linter.CheckerContext
+	engine *ruleguard.Engine
+}
+
+func (c *embeddedRuleguardChecker) WalkFile(f *ast.File) {
+	runRuleguardEngine(c.ctx, f, c.engine, &ruleguard.RunContext{
+		Pkg:   c.ctx.Pkg,
+		Types: c.ctx.TypesInfo,
+		Sizes: c.ctx.SizesInfo,
+		Fset:  c.ctx.FileSet,
+	})
+}