diff --git a/pkg/detectors/cloudinary/cloudinary.go b/pkg/detectors/cloudinary/cloudinary.go new file mode 100644 index 000000000..6db54f3af --- /dev/null +++ b/pkg/detectors/cloudinary/cloudinary.go @@ -0,0 +1,134 @@ +package cloudinary + +import ( + "context" + "fmt" + "io" + "net/http" + "net/url" + "path" + + regexp "github.com/wasilibs/go-re2" + + "github.com/trufflesecurity/trufflehog/v3/pkg/common" + "github.com/trufflesecurity/trufflehog/v3/pkg/detectors" + "github.com/trufflesecurity/trufflehog/v3/pkg/pb/detector_typepb" +) + +type Scanner struct { + client *http.Client + detectors.DefaultMultiPartCredentialProvider +} + +// Ensure the Scanner satisfies the interface at compile time. +var _ detectors.Detector = (*Scanner)(nil) + +var ( + defaultClient = common.SaneHttpClient() + + // Cloudinary cloud names are typically b/w 3-50 characters. + // This regex matches cloud names that appear near the word "cloudinary" and either the keyword "name" or "@" (as in URLs). + cloudnamePat = regexp.MustCompile(`(?i:cloudinary)(?:.|[\n\r]){0,47}?` + `(?i:name|@)(?:.){0,10}?` + `\b([a-zA-Z][a-zA-Z0-9-]{2,49})\b`) + + // Cloudinary API keys are numeric and typically 15 digits long. + apiKeyPat = regexp.MustCompile(detectors.PrefixRegex([]string{"cloudinary"}) + `\b(\d{15})\b`) + + // Cloudinary API secrets are typically 27 characters long and may contain + // uppercase letters, lowercase letters, digits, underscores, or hyphens. + apiSecretPat = regexp.MustCompile(detectors.PrefixRegex([]string{"cloudinary"}) + `\b([A-Za-z0-9_-]{27})\b`) +) + +// Keywords are used for efficiently pre-filtering chunks. +// Use identifiers in the secret preferably, or the provider name. +func (s Scanner) Keywords() []string { + return []string{"cloudinary"} +} + +func (s Scanner) getClient() *http.Client { + if s.client != nil { + return s.client + } + return defaultClient +} + +// FromData will find and optionally verify DeepSeek secrets in a given set of bytes. +func (s Scanner) FromData(ctx context.Context, verify bool, data []byte) (results []detectors.Result, err error) { + dataStr := string(data) + + uniqueCloudNames := make(map[string]struct{}) + uniqueApiKeys := make(map[string]struct{}) + uniqueApiSecret := make(map[string]struct{}) + for _, match := range cloudnamePat.FindAllStringSubmatch(dataStr, -1) { + uniqueCloudNames[match[1]] = struct{}{} + } + for _, match := range apiKeyPat.FindAllStringSubmatch(dataStr, -1) { + uniqueApiKeys[match[1]] = struct{}{} + } + for _, match := range apiSecretPat.FindAllStringSubmatch(dataStr, -1) { + uniqueApiSecret[match[1]] = struct{}{} + } + for cloudName := range uniqueCloudNames { + for apiKey := range uniqueApiKeys { + for apiSecret := range uniqueApiSecret { + s1 := detectors.Result{ + DetectorType: detector_typepb.DetectorType_Cloudinary, + Raw: []byte(apiKey), + RawV2: []byte(fmt.Sprintf("%s:%s:%s", cloudName, apiKey, apiSecret)), + SecretParts: map[string]string{ + "cloud_name": cloudName, + "api_key": apiKey, + "api_secret": apiSecret, + }, + } + if verify { + verified, verificationErr := verifyToken(ctx, s.getClient(), cloudName, apiKey, apiSecret) + s1.SetVerificationError(verificationErr, cloudName, apiKey, apiSecret) + s1.Verified = verified + } + results = append(results, s1) + } + } + } + return +} + +func verifyToken(ctx context.Context, client *http.Client, cloudName, apiKey, apiSecret string) (bool, error) { + u := &url.URL{ + Scheme: "https", + Host: "api.cloudinary.com", + Path: path.Join("v1_1", cloudName, "usage"), + } + req, err := http.NewRequestWithContext(ctx, http.MethodGet, u.String(), http.NoBody) + if err != nil { + return false, err + } + + req.SetBasicAuth(apiKey, apiSecret) + + res, err := client.Do(req) + if err != nil { + return false, err + } + defer func() { + _, _ = io.Copy(io.Discard, res.Body) + _ = res.Body.Close() + }() + + switch res.StatusCode { + case http.StatusOK: + return true, nil + case http.StatusUnauthorized: + // Invalid + return false, nil + default: + return false, fmt.Errorf("unexpected HTTP response status %d", res.StatusCode) + } +} + +func (s Scanner) Type() detector_typepb.DetectorType { + return detector_typepb.DetectorType_Cloudinary +} + +func (s Scanner) Description() string { + return "Cloudinary is a cloud-based media management platform. Cloudinary API keys can be used to upload, manage, and deliver media assets." +} diff --git a/pkg/detectors/cloudinary/cloudinary_integration_test.go b/pkg/detectors/cloudinary/cloudinary_integration_test.go new file mode 100644 index 000000000..04d2b0c72 --- /dev/null +++ b/pkg/detectors/cloudinary/cloudinary_integration_test.go @@ -0,0 +1,168 @@ +//go:build detectors +// +build detectors + +package cloudinary + +import ( + "context" + "fmt" + "testing" + "time" + + "github.com/google/go-cmp/cmp" + "github.com/google/go-cmp/cmp/cmpopts" + + "github.com/trufflesecurity/trufflehog/v3/pkg/common" + "github.com/trufflesecurity/trufflehog/v3/pkg/detectors" + "github.com/trufflesecurity/trufflehog/v3/pkg/pb/detector_typepb" +) + +func TestCloudinary_FromChunk(t *testing.T) { + ctx, cancel := context.WithTimeout(context.Background(), time.Second*5) + defer cancel() + testSecrets, err := common.GetSecret(ctx, "trufflehog-testing", "detectors6") + if err != nil { + t.Fatalf("could not get test secrets from GCP: %s", err) + } + cloudName := testSecrets.MustGetField("CLOUDINARY_CLOUD_NAME") + apiKey := testSecrets.MustGetField("CLOUDINARY_API_KEY") + apiSecret := testSecrets.MustGetField("CLOUDINARY_SECRET_KEY") + inactiveSecret := testSecrets.MustGetField("CLOUDINARY_INACTIVE_SECRET_KEY") + + type args struct { + ctx context.Context + data []byte + verify bool + } + tests := []struct { + name string + s Scanner + args args + want []detectors.Result + wantErr bool + wantVerificationErr bool + }{ + { + name: "found, verified", + s: Scanner{}, + args: args{ + ctx: context.Background(), + data: fmt.Appendf([]byte{}, "You can find a Cloudinary apiSecret %s, Cloudinary apiKey %v and Cloudinary cloudName %v", apiSecret, apiKey, cloudName), + verify: true, + }, + want: []detectors.Result{ + { + DetectorType: detector_typepb.DetectorType_Cloudinary, + Verified: true, + Raw: []byte(apiKey), + RawV2: []byte(fmt.Sprintf("%s:%s:%s", cloudName, apiKey, apiSecret)), + }, + }, + wantErr: false, + }, + { + name: "found, real secrets, verification error due to timeout", + s: Scanner{client: common.SaneHttpClientTimeOut(1 * time.Microsecond)}, + args: args{ + ctx: context.Background(), + data: fmt.Appendf([]byte{}, "You can find a Cloudinary apiSecret %s, Cloudinary apiKey %v and Cloudinary cloudName %v", apiSecret, apiKey, cloudName), + verify: true, + }, + want: []detectors.Result{ + { + DetectorType: detector_typepb.DetectorType_Cloudinary, + Verified: false, + Raw: []byte(apiKey), + RawV2: []byte(fmt.Sprintf("%s:%s:%s", cloudName, apiKey, apiSecret)), + }, + }, + wantErr: false, + wantVerificationErr: true, + }, + { + name: "found, real secrets, verification error due to unexpected api surface", + s: Scanner{client: common.ConstantResponseHttpClient(500, "{}")}, + args: args{ + ctx: context.Background(), + data: fmt.Appendf([]byte{}, "You can find a Cloudinary apiSecret %s, Cloudinary apiKey %v and Cloudinary cloudName %v", apiSecret, apiKey, cloudName), + verify: true, + }, + want: []detectors.Result{ + { + DetectorType: detector_typepb.DetectorType_Cloudinary, + Verified: false, + Raw: []byte(apiKey), + RawV2: []byte(fmt.Sprintf("%s:%s:%s", cloudName, apiKey, apiSecret)), + }, + }, + wantErr: false, + wantVerificationErr: true, + }, + { + name: "found, unverified", + s: Scanner{}, + args: args{ + ctx: context.Background(), + data: fmt.Appendf([]byte{}, "You can find a Cloudinary inactiveapiSecret %s, Cloudinary apiKey %v and Cloudinary cloudName %v", inactiveSecret, apiKey, cloudName), + verify: true, + }, + want: []detectors.Result{ + { + DetectorType: detector_typepb.DetectorType_Cloudinary, + Verified: false, + Raw: []byte(apiKey), + RawV2: []byte(fmt.Sprintf("%s:%s:%s", cloudName, apiKey, inactiveSecret)), + }, + }, + wantErr: false, + }, + { + name: "not found", + s: Scanner{}, + args: args{ + ctx: context.Background(), + data: []byte("You cannot find the secret within"), + verify: true, + }, + want: nil, + wantErr: false, + }, + } + for _, tt := range tests { + t.Run(tt.name, func(t *testing.T) { + got, err := tt.s.FromData(tt.args.ctx, tt.args.verify, tt.args.data) + if (err != nil) != tt.wantErr { + t.Errorf("Cloudinary.FromData() error = %v, wantErr %v", err, tt.wantErr) + return + } + for i := range got { + if len(got[i].Raw) == 0 { + t.Fatal("no raw secret present") + } + if (got[i].VerificationError() != nil) != tt.wantVerificationErr { + t.Fatalf("wantVerificationError = %v, verification error = %v", tt.wantVerificationErr, got[i].VerificationError()) + } + } + ignoreOpts := cmpopts.IgnoreFields(detectors.Result{}, "ExtraData", "verificationError", "primarySecret", "SecretParts") + if diff := cmp.Diff(got, tt.want, ignoreOpts); diff != "" { + t.Errorf("Cloudinary.FromData() %s diff: (-got +want)\n%s", tt.name, diff) + } + }) + } +} + +func BenchmarkFromData(benchmark *testing.B) { + ctx := context.Background() + s := Scanner{} + for name, data := range detectors.MustGetBenchmarkData() { + benchmark.Run(name, func(b *testing.B) { + b.ResetTimer() + for n := 0; n < b.N; n++ { + _, err := s.FromData(ctx, false, data) + if err != nil { + b.Fatal(err) + } + } + }) + } +} diff --git a/pkg/detectors/cloudinary/cloudinary_test.go b/pkg/detectors/cloudinary/cloudinary_test.go new file mode 100644 index 000000000..7f3f2a828 --- /dev/null +++ b/pkg/detectors/cloudinary/cloudinary_test.go @@ -0,0 +1,151 @@ +package cloudinary + +import ( + "context" + "testing" + + "github.com/google/go-cmp/cmp" + "github.com/stretchr/testify/require" + + "github.com/trufflesecurity/trufflehog/v3/pkg/detectors" + "github.com/trufflesecurity/trufflehog/v3/pkg/engine/ahocorasick" +) + +func TestCloudinary_Pattern(t *testing.T) { + d := Scanner{} + ahoCorasickCore := ahocorasick.NewAhoCorasickCore([]detectors.Detector{d}) + + tests := []struct { + name string + input string + want []string + }{ + { + name: "valid pattern", + input: ` + [INFO] Sending request to the API + [DEBUG] Using cloudinary_apiSecret=ndiawudh1_wdajwoidjajawdeps + [DEBUG] Using cloudinary_apiKey=218873249723411 + [DEBUG] Using cloudinary_cloudName=wdjaiwojd + [INFO] Response received: 200 OK + `, + want: []string{"wdjaiwojd" + ":" + "218873249723411" + ":" + "ndiawudh1_wdajwoidjajawdeps"}, + }, + { + name: "valid pattern - cloudinary api environment variable", + input: ` + [INFO] Sending request to the cloudinary + [DEBUG] Using url=cloudinary://715268876851676:V-fwwRhcp3VrRPAqaFkLq3rpa60@fakwfpoaj + [INFO] Response received: 200 OK + `, + want: []string{"fakwfpoaj" + ":" + "715268876851676" + ":" + "V-fwwRhcp3VrRPAqaFkLq3rpa60"}, + }, + + { + name: "valid pattern - out of prefix range - apikey", + input: ` + [INFO] Sending request to the cloudinary + [DEBUG] Using cloudName=wdjaiwojd + [DEBUG] Using cloudinary_apiSecret=ndiawudh1_wdajwoidjajawdeps + [DEBUG] apiKey=218873249723411 + [INFO] Response received: 200 OK + `, + want: nil, + }, + { + name: "valid pattern - out of prefix range - apiSecret", + input: ` + [INFO] Sending request to the cloudinary + [DEBUG] Using cloudName=wdjaiwojd + [DEBUG] apiKey=218873249723411 + [DEBUG] Using apiSecret=ndiawudh1_wdajwoidjajawdeps + [INFO] Response received: 200 OK + `, + want: nil, + }, + { + name: "valid pattern - out of prefix range - cloudName", + input: ` + [INFO] Sending request to the cloudinary + [DEBUG] apiKey=218873249723411 + [DEBUG] Using apiSecret=ndiawudh1_wdajwoidjajawdeps + [INFO] Response received: 200 OK + [DEBUG] Used cloudName=wdjaiwojd + `, + want: nil, + }, + { + name: "valid pattern - only apikey", + input: ` + [INFO] Sending request to the cloudinary API + [DEBUG] Using apiKey=218873249723411 + `, + want: nil, + }, + { + name: "valid pattern - only secret", + input: ` + [INFO] Sending request to the cloudinary API + [DEBUG] Using apiSecret=ndiawudh1_wdajwoidjajawdeps + [INFO] Response received: 200 OK + `, + want: nil, + }, + { + name: "valid pattern - only cloudName", + input: ` + [INFO] Sending request to the cloudinary API + [DEBUG] Using cloudName=wdjaiwojd + [INFO] Response received: 200 OK + `, + want: nil, + }, + { + name: "invalid pattern - invalid api key length", + input: ` + [INFO] Sending request to the cloudinary API + [DEBUG] Using apikey=12312312432541444 + [DEBUG] Using apiSecret=ndiawudh1_wdajwoidjajawdeps + [DEBUG] Using cloudName=wdjaiwojd + [ERROR] Response received: 400 BadRequest + `, + want: nil, + }, + } + + for _, test := range tests { + t.Run(test.name, func(t *testing.T) { + matchedDetectors := ahoCorasickCore.FindDetectorMatches([]byte(test.input)) + if len(matchedDetectors) == 0 { + t.Errorf("test %q failed: expected keywords %v to be found in the input", test.name, d.Keywords()) + return + } + + results, err := d.FromData(context.Background(), false, []byte(test.input)) + require.NoError(t, err) + + if len(results) != len(test.want) { + t.Errorf("mismatch in result count: expected %d, got %d", len(test.want), len(results)) + return + } + + actual := make(map[string]struct{}, len(results)) + for _, r := range results { + if len(r.RawV2) > 0 { + actual[string(r.RawV2)] = struct{}{} + } else { + actual[string(r.Raw)] = struct{}{} + } + } + + expected := make(map[string]struct{}, len(test.want)) + for _, v := range test.want { + expected[v] = struct{}{} + } + + if diff := cmp.Diff(expected, actual); diff != "" { + t.Errorf("%s diff: (-want +got)\n%s", test.name, diff) + } + }) + } +} diff --git a/pkg/engine/defaults/defaults.go b/pkg/engine/defaults/defaults.go index f769af2c5..46db6f158 100644 --- a/pkg/engine/defaults/defaults.go +++ b/pkg/engine/defaults/defaults.go @@ -165,6 +165,7 @@ import ( "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/cloudflarecakey" "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/cloudflareglobalapikey" "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/cloudimage" + "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/cloudinary" "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/cloudmersive" "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/cloudplan" "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/cloudsmith" @@ -1037,6 +1038,7 @@ func buildDetectorList() []detectors.Detector { &cloudflarecakey.Scanner{}, &cloudflareglobalapikey.Scanner{}, &cloudimage.Scanner{}, + &cloudinary.Scanner{}, &cloudmersive.Scanner{}, &cloudplan.Scanner{}, &cloudsmith.Scanner{}, diff --git a/pkg/pb/detector_typepb/detector_type.pb.go b/pkg/pb/detector_typepb/detector_type.pb.go index 51f668528..9876452be 100644 --- a/pkg/pb/detector_typepb/detector_type.pb.go +++ b/pkg/pb/detector_typepb/detector_type.pb.go @@ -1101,6 +1101,7 @@ const ( DetectorType_BitbucketDataCenter DetectorType = 1045 DetectorType_JiraDataCenterPAT DetectorType = 1046 DetectorType_ConfluenceDataCenter DetectorType = 1047 + DetectorType_Cloudinary DetectorType = 1048 ) // Enum value maps for DetectorType. @@ -2150,6 +2151,7 @@ var ( 1045: "BitbucketDataCenter", 1046: "JiraDataCenterPAT", 1047: "ConfluenceDataCenter", + 1048: "Cloudinary", } DetectorType_value = map[string]int32{ "Alibaba": 0, @@ -3196,6 +3198,7 @@ var ( "BitbucketDataCenter": 1045, "JiraDataCenterPAT": 1046, "ConfluenceDataCenter": 1047, + "Cloudinary": 1048, } ) @@ -3231,7 +3234,7 @@ var File_detector_type_proto protoreflect.FileDescriptor var file_detector_type_proto_rawDesc = []byte{ 0x0a, 0x13, 0x64, 0x65, 0x74, 0x65, 0x63, 0x74, 0x6f, 0x72, 0x5f, 0x74, 0x79, 0x70, 0x65, 0x2e, 0x70, 0x72, 0x6f, 0x74, 0x6f, 0x12, 0x0d, 0x64, 0x65, 0x74, 0x65, 0x63, 0x74, 0x6f, 0x72, 0x5f, - 0x74, 0x79, 0x70, 0x65, 0x2a, 0x8c, 0x88, 0x01, 0x0a, 0x0c, 0x44, 0x65, 0x74, 0x65, 0x63, 0x74, + 0x74, 0x79, 0x70, 0x65, 0x2a, 0x9d, 0x88, 0x01, 0x0a, 0x0c, 0x44, 0x65, 0x74, 0x65, 0x63, 0x74, 0x6f, 0x72, 0x54, 0x79, 0x70, 0x65, 0x12, 0x0b, 0x0a, 0x07, 0x41, 0x6c, 0x69, 0x62, 0x61, 0x62, 0x61, 0x10, 0x00, 0x12, 0x08, 0x0a, 0x04, 0x41, 0x4d, 0x51, 0x50, 0x10, 0x01, 0x12, 0x07, 0x0a, 0x03, 0x41, 0x57, 0x53, 0x10, 0x02, 0x12, 0x09, 0x0a, 0x05, 0x41, 0x7a, 0x75, 0x72, 0x65, 0x10, @@ -4320,11 +4323,12 @@ var file_detector_type_proto_rawDesc = []byte{ 0x08, 0x12, 0x16, 0x0a, 0x11, 0x4a, 0x69, 0x72, 0x61, 0x44, 0x61, 0x74, 0x61, 0x43, 0x65, 0x6e, 0x74, 0x65, 0x72, 0x50, 0x41, 0x54, 0x10, 0x96, 0x08, 0x12, 0x19, 0x0a, 0x14, 0x43, 0x6f, 0x6e, 0x66, 0x6c, 0x75, 0x65, 0x6e, 0x63, 0x65, 0x44, 0x61, 0x74, 0x61, 0x43, 0x65, 0x6e, 0x74, 0x65, - 0x72, 0x10, 0x97, 0x08, 0x42, 0x41, 0x5a, 0x3f, 0x67, 0x69, 0x74, 0x68, 0x75, 0x62, 0x2e, 0x63, - 0x6f, 0x6d, 0x2f, 0x74, 0x72, 0x75, 0x66, 0x66, 0x6c, 0x65, 0x73, 0x65, 0x63, 0x75, 0x72, 0x69, - 0x74, 0x79, 0x2f, 0x74, 0x72, 0x75, 0x66, 0x66, 0x6c, 0x65, 0x68, 0x6f, 0x67, 0x2f, 0x76, 0x33, - 0x2f, 0x70, 0x6b, 0x67, 0x2f, 0x70, 0x62, 0x2f, 0x64, 0x65, 0x74, 0x65, 0x63, 0x74, 0x6f, 0x72, - 0x5f, 0x74, 0x79, 0x70, 0x65, 0x70, 0x62, 0x62, 0x06, 0x70, 0x72, 0x6f, 0x74, 0x6f, 0x33, + 0x72, 0x10, 0x97, 0x08, 0x12, 0x0f, 0x0a, 0x0a, 0x43, 0x6c, 0x6f, 0x75, 0x64, 0x69, 0x6e, 0x61, + 0x72, 0x79, 0x10, 0x98, 0x08, 0x42, 0x41, 0x5a, 0x3f, 0x67, 0x69, 0x74, 0x68, 0x75, 0x62, 0x2e, + 0x63, 0x6f, 0x6d, 0x2f, 0x74, 0x72, 0x75, 0x66, 0x66, 0x6c, 0x65, 0x73, 0x65, 0x63, 0x75, 0x72, + 0x69, 0x74, 0x79, 0x2f, 0x74, 0x72, 0x75, 0x66, 0x66, 0x6c, 0x65, 0x68, 0x6f, 0x67, 0x2f, 0x76, + 0x33, 0x2f, 0x70, 0x6b, 0x67, 0x2f, 0x70, 0x62, 0x2f, 0x64, 0x65, 0x74, 0x65, 0x63, 0x74, 0x6f, + 0x72, 0x5f, 0x74, 0x79, 0x70, 0x65, 0x70, 0x62, 0x62, 0x06, 0x70, 0x72, 0x6f, 0x74, 0x6f, 0x33, } var ( diff --git a/proto/detector_type.proto b/proto/detector_type.proto index ddd4c83c6..5b0f0f42f 100644 --- a/proto/detector_type.proto +++ b/proto/detector_type.proto @@ -1049,4 +1049,5 @@ enum DetectorType { BitbucketDataCenter = 1045; JiraDataCenterPAT = 1046; ConfluenceDataCenter = 1047; + Cloudinary = 1048; }