adding mime type file support

v2/go.mod

@@ -114,6 +114,7 @@ require (
 	github.com/google/uuid v1.3.0 // indirect
 	github.com/gosuri/uilive v0.0.4 // indirect
 	github.com/gosuri/uiprogress v0.0.1 // indirect
+	github.com/h2non/filetype v1.1.3 // indirect
 	github.com/hashicorp/go-cleanhttp v0.5.1 // indirect
 	github.com/hashicorp/go-retryablehttp v0.6.8 // indirect
 	github.com/iancoleman/orderedmap v0.0.0-20190318233801-ac98e3ecb4b0 // indirect

v2/go.sum

@@ -263,6 +263,8 @@ github.com/gosuri/uilive v0.0.4 h1:hUEBpQDj8D8jXgtCdBu7sWsy5sbW/5GhuO8KBwJ2jyY=
 github.com/gosuri/uilive v0.0.4/go.mod h1:V/epo5LjjlDE5RJUcqx8dbw+zc93y5Ya3yg8tfZ74VI=
 github.com/gosuri/uiprogress v0.0.1 h1:0kpv/XY/qTmFWl/SkaJykZXrBBzwwadmW8fRb7RJSxw=
 github.com/gosuri/uiprogress v0.0.1/go.mod h1:C1RTYn4Sc7iEyf6j8ft5dyoZ4212h8G1ol9QQluh5+0=
+github.com/h2non/filetype v1.1.3 h1:FKkx9QbD7HR/zjK1Ia5XiBsq9zdLi5Kf3zGyFTAFkGg=
+github.com/h2non/filetype v1.1.3/go.mod h1:319b3zT68BvV+WRj7cwy856M2ehB3HqNOt6sy1HndBY=
 github.com/hashicorp/go-cleanhttp v0.5.1 h1:dH3aiDG9Jvb5r5+bYHsikaOUIpcM0xvgMXVoDkXMzJM=
 github.com/hashicorp/go-cleanhttp v0.5.1/go.mod h1:JpRdi6/HCYpAwUzNwuwqhbovhLtngrth3wmdIIUrZ80=
 github.com/hashicorp/go-hclog v0.9.2 h1:CG6TE5H9/JXsFWJCfoIVpKFIkFe6ysEuHirp4DxCsHI=

v2/pkg/protocols/file/file.go

@@ -5,6 +5,7 @@ import (
 	"strings"
 
 	"github.com/docker/go-units"
+	"github.com/h2non/filetype"
 	"github.com/pkg/errors"
 
 	"github.com/projectdiscovery/nuclei/v2/pkg/operators"
@@ -21,12 +22,12 @@ type Request struct {
 	// Operators for the current request go here.
 	operators.Operators `yaml:",inline"`
 	// description: |
-	//   Extensions is the list of extensions to perform matching on.
+	//   Extensions is the list of extensions or mime types to perform matching on.
 	// examples:
 	//   - value: '[]string{".txt", ".go", ".json"}'
 	Extensions []string `yaml:"extensions,omitempty" jsonschema:"title=extensions to match,description=List of extensions to perform matching on"`
 	// description: |
-	//   DenyList is the list of file, directories or extensions to deny during matching.
+	//   DenyList is the list of file, directories, mime types or extensions to deny during matching.
 	//
 	//   By default, it contains some non-interesting extensions that are hardcoded
 	//   in nuclei.
@@ -55,9 +56,11 @@ type Request struct {
 	CompiledOperators *operators.Operators `yaml:"-"`
 
 	// cache any variables that may be needed for operation.
-	options    *protocols.ExecuterOptions
-	extensions map[string]struct{}
-	denyList   map[string]struct{}
+	options             *protocols.ExecuterOptions
+	mimeTypesChecks     []string
+	extensions          map[string]struct{}
+	denyList            map[string]struct{}
+	denyMimeTypesChecks []string
 
 	// description: |
 	//   NoRecursive specifies whether to not do recursive checks if folders are provided.
@@ -120,15 +123,20 @@ func (request *Request) Compile(options *protocols.ExecuterOptions) error {
 	request.denyList = make(map[string]struct{})
 
 	for _, extension := range request.Extensions {
-		if extension == "all" {
+		switch {
+		case extension == "all":
 			request.allExtensions = true
-		} else {
+		case filetype.IsMIMESupported(extension):
+			continue
+		default:
 			if !strings.HasPrefix(extension, ".") {
 				extension = "." + extension
 			}
 			request.extensions[extension] = struct{}{}
 		}
 	}
+	request.mimeTypesChecks = extractMimeTypes(request.Extensions)
+
 	// process default denylist (extensions)
 	var denyList []string
 	if !request.Archive {
@@ -147,9 +155,30 @@ func (request *Request) Compile(options *protocols.ExecuterOptions) error {
 		// also add a cleaned version as the exclusion path can be dirty (eg. /a/b/c, /a/b/c/, a///b///c/../d)
 		request.denyList[filepath.Clean(excludeItem)] = struct{}{}
 	}
+	request.denyMimeTypesChecks = extractMimeTypes(request.DenyList)
 	return nil
 }
 
+func matchAnyMimeTypes(data []byte, mimeTypes []string) bool {
+	for _, mimeType := range mimeTypes {
+		if filetype.Is(data, mimeType) {
+			return true
+		}
+	}
+	return false
+}
+
+func extractMimeTypes(m []string) []string {
+	var mimeTypes []string
+	for _, mm := range m {
+		if !filetype.IsMIMESupported(mm) {
+			continue
+		}
+		mimeTypes = append(mimeTypes, mm)
+	}
+	return mimeTypes
+}
+
 // Requests returns the total number of requests the YAML rule will perform
 func (request *Request) Requests() int {
 	return 0

v2/pkg/protocols/file/find.go

@@ -1,12 +1,14 @@
 package file
 
 import (
+	"io"
 	"os"
 	"path/filepath"
 	"strings"
 
 	"github.com/karrick/godirwalk"
 	"github.com/pkg/errors"
+	"github.com/projectdiscovery/fileutil"
 	"github.com/projectdiscovery/folderutil"
 	"github.com/projectdiscovery/gologger"
 )
@@ -109,7 +111,7 @@ func (request *Request) findDirectoryMatches(absPath string, processed map[strin
 // validatePath validates a file path for blacklist and whitelist options
 func (request *Request) validatePath(absPath, item string) bool {
 	extension := filepath.Ext(item)
-
+	// extension check
 	if len(request.extensions) > 0 {
 		if _, ok := request.extensions[extension]; ok {
 			return true
@@ -117,11 +119,30 @@ func (request *Request) validatePath(absPath, item string) bool {
 			return false
 		}
 	}
+
+	// mime type check
+	// read first bytes to infer runtime type
+	fileExists := fileutil.FileExists(item)
+	var dataChunk []byte
+	if fileExists {
+		dataChunk, _ = readChunk(item)
+		if len(request.mimeTypesChecks) > 0 && matchAnyMimeTypes(dataChunk, request.mimeTypesChecks) {
+			return true
+		}
+	}
+
 	if matchingRule, ok := request.isInDenyList(absPath, item); ok {
 		gologger.Verbose().Msgf("Ignoring path %s due to denylist item %s\n", item, matchingRule)
 		return false
 	}
 
+	// denied mime type checks
+	if fileExists {
+		if len(request.denyMimeTypesChecks) > 0 && matchAnyMimeTypes(dataChunk, request.denyMimeTypesChecks) {
+			return false
+		}
+	}
+
 	return true
 }
 
@@ -175,6 +196,21 @@ func (request *Request) isInDenyList(absPath, item string) (string, bool) {
 	return "", false
 }
 
+func readChunk(fileName string) ([]byte, error) {
+	r, err := os.Open(fileName)
+	if err != nil {
+		return nil, err
+	}
+
+	defer r.Close()
+
+	var buff [1024]byte
+	if _, err = io.ReadFull(r, buff[:]); err != nil {
+		return nil, err
+	}
+	return buff[:], nil
+}
+
 func (request *Request) isAnyChunkInDenyList(path string, splitWithUtils bool) (string, bool) {
 	var paths []string