[INS-377] Add V2 for weights and biases detector (#4951)

Co-authored-by: steeeve <[email protected]>

main.go

@@ -575,6 +575,7 @@ func run(state overseer.State, logSync func() error) {
 	feature.NewRelicMobileAppTokenDetectorEnabled.Store(true)
 	feature.MSTeamsWebhookV2DetectorEnabled.Store(true)
 	feature.SolarwindsDetectorEnabled.Store(true)
+	feature.WeightsAndBiasesV2DetectorEnabled.Store(true)
 
 	conf := &config.Config{}
 	if *configFilename != "" {

pkg/detectors/weightsandbiases/base.go

@@ -17,53 +17,46 @@ import (
 	"github.com/trufflesecurity/trufflehog/v3/pkg/pb/detector_typepb"
 )
 
-type Scanner struct{ client *http.Client }
+// BaseScanner is a base struct embedded by versioned scanners. It holds the HTTP client and
+// shared detection/verification logic.
+type BaseScanner struct {
+	Client *http.Client
+}
 
-// Ensure the Scanner satisfies the interface at compile time.
-var _ detectors.Detector = (*Scanner)(nil)
+var defaultClient = common.SaneHttpClient()
 
-var (
-	defaultClient = common.SaneHttpClient()
-	// Make sure that your group is surrounded in boundary characters such as below to reduce false positives.
-	keyPat = regexp.MustCompile(detectors.PrefixRegex([]string{"wandb"}) + `\b([0-9a-f]{40})\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{"wandb"} }
-
-// FromData will find and optionally verify Weightsandbiases secrets in a given set of bytes.
-func (s Scanner) FromData(ctx context.Context, verify bool, data []byte) (results []detectors.Result, err error) {
+// FromData finds and optionally verifies WeightsAndBiases secrets in data using the provided
+// pattern. version is included in ExtraData of each result.
+func (s BaseScanner) FromData(ctx context.Context, verify bool, data []byte, keyPat *regexp.Regexp, version int) ([]detectors.Result, error) {
 	dataStr := string(data)
-
 	uniqueMatches := make(map[string]struct{})
 	for _, match := range keyPat.FindAllStringSubmatch(dataStr, -1) {
 		uniqueMatches[match[1]] = struct{}{}
 	}
 
+	var results []detectors.Result
 	for match := range uniqueMatches {
-		s1 := detectors.Result{
+		r := detectors.Result{
 			DetectorType: detector_typepb.DetectorType_WeightsAndBiases,
 			Raw:          []byte(match),
 			SecretParts:  map[string]string{"key": match},
 		}
 
 		if verify {
-			client := s.client
-			if client == nil {
-				client = defaultClient
-			}
-
-			isVerified, extraData, verificationErr := verifyMatch(ctx, client, match)
-			s1.Verified = isVerified
-			s1.ExtraData = extraData
-			s1.SetVerificationError(verificationErr, match)
+			isVerified, extraData, verificationErr := s.verifyMatch(ctx, match)
+			r.Verified = isVerified
+			r.ExtraData = extraData
+			r.SetVerificationError(verificationErr, match)
 		}
 
-		results = append(results, s1)
-	}
+		if r.ExtraData == nil {
+			r.ExtraData = make(map[string]string)
+		}
+		r.ExtraData["version"] = strconv.Itoa(version)
 
-	return
+		results = append(results, r)
+	}
+	return results, nil
 }
 
 type viewerResponse struct {
@@ -77,7 +70,16 @@ type viewerResponse struct {
 	} `json:"data"`
 }
 
-func verifyMatch(ctx context.Context, client *http.Client, token string) (bool, map[string]string, error) {
+// verifyMatch checks the credential against the W&B GraphQL /graphql endpoint using the viewer query,
+// which requires no special permissions. A 200 with a non-empty username means the token is valid;
+// 401 means invalid or revoked.
+// Docs: https://docs.wandb.ai/ref/graphql
+func (s BaseScanner) verifyMatch(ctx context.Context, token string) (bool, map[string]string, error) {
+	client := s.Client
+	if client == nil {
+		client = defaultClient
+	}
+
 	query := `{"query": "query Viewer { viewer { id username email admin } }"}`
 
 	const baseURL = "https://api.wandb.ai/graphql"
@@ -124,10 +126,10 @@ func verifyMatch(ctx context.Context, client *http.Client, token string) (bool,
 	}
 }
 
-func (s Scanner) Description() string {
+func (s BaseScanner) Description() string {
 	return "Weights & Biases is a Machine Learning Operations (MLOps) platform that helps track experiments, version datasets, evaluate model performance, and collaborate with team members"
 }
 
-func (s Scanner) Type() detector_typepb.DetectorType {
+func (s BaseScanner) Type() detector_typepb.DetectorType {
 	return detector_typepb.DetectorType_WeightsAndBiases
 }

pkg/detectors/weightsandbiases/v1/weightsandbiases.go

@@ -0,0 +1,29 @@
+package weightsandbiases
+
+import (
+	"context"
+
+	regexp "github.com/wasilibs/go-re2"
+
+	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors"
+	base "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/weightsandbiases"
+)
+
+type Scanner struct {
+	base.BaseScanner
+}
+
+// Ensure the Scanner satisfies the interface at compile time.
+var _ detectors.Detector = (*Scanner)(nil)
+var _ detectors.Versioner = (*Scanner)(nil)
+
+var keyPat = regexp.MustCompile(detectors.PrefixRegex([]string{"wandb"}) + `\b([0-9a-f]{40})\b`)
+
+func (s Scanner) Version() int { return 1 }
+
+func (s Scanner) Keywords() []string { return []string{"wandb"} }
+
+func (s Scanner) FromData(ctx context.Context, verify bool, data []byte) ([]detectors.Result, error) {
+	return s.BaseScanner.FromData(ctx, verify, data, keyPat, s.Version())
+}
+

pkg/detectors/weightsandbiases/v1/weightsandbiases_integration_test.go

@@ -14,6 +14,7 @@ import (
 
 	"github.com/trufflesecurity/trufflehog/v3/pkg/common"
 	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors"
+	base "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/weightsandbiases"
 	"github.com/trufflesecurity/trufflehog/v3/pkg/pb/detector_typepb"
 )
 
@@ -56,6 +57,7 @@ func TestWeightsandbiases_FromChunk(t *testing.T) {
 						"admin":    "false",
 						"email":    "[email protected]",
 						"username": "source-integrations",
+						"version":  "1",
 					},
 				},
 			},
@@ -74,6 +76,7 @@ func TestWeightsandbiases_FromChunk(t *testing.T) {
 				{
 					DetectorType: detector_typepb.DetectorType_WeightsAndBiases,
 					Verified:     false,
+					ExtraData:    map[string]string{"version": "1"},
 				},
 			},
 			wantErr:             false,
@@ -93,7 +96,7 @@ func TestWeightsandbiases_FromChunk(t *testing.T) {
 		},
 		{
 			name: "found, would be verified if not for timeout",
-			s:    Scanner{client: common.SaneHttpClientTimeOut(1 * time.Microsecond)},
+			s:    Scanner{BaseScanner: base.BaseScanner{Client: common.SaneHttpClientTimeOut(1 * time.Microsecond)}},
 			args: args{
 				ctx:    context.Background(),
 				data:   []byte(fmt.Sprintf("You can find a weightsandbiases secret wandb %s within", secret)),
@@ -103,6 +106,7 @@ func TestWeightsandbiases_FromChunk(t *testing.T) {
 				{
 					DetectorType: detector_typepb.DetectorType_WeightsAndBiases,
 					Verified:     false,
+					ExtraData:    map[string]string{"version": "1"},
 				},
 			},
 			wantErr:             false,
@@ -110,7 +114,7 @@ func TestWeightsandbiases_FromChunk(t *testing.T) {
 		},
 		{
 			name: "found, verified but unexpected api surface",
-			s:    Scanner{client: common.ConstantResponseHttpClient(404, "")},
+			s:    Scanner{BaseScanner: base.BaseScanner{Client: common.ConstantResponseHttpClient(404, "")}},
 			args: args{
 				ctx:    context.Background(),
 				data:   []byte(fmt.Sprintf("You can find a weightsandbiases secret wandb %s within", secret)),
@@ -120,6 +124,7 @@ func TestWeightsandbiases_FromChunk(t *testing.T) {
 				{
 					DetectorType: detector_typepb.DetectorType_WeightsAndBiases,
 					Verified:     false,
+					ExtraData:    map[string]string{"version": "1"},
 				},
 			},
 			wantErr:             false,
@@ -138,11 +143,14 @@ func TestWeightsandbiases_FromChunk(t *testing.T) {
 				if len(got[i].Raw) == 0 {
 					t.Fatalf("no raw secret present: \n %+v", got[i])
 				}
+				if len(got[i].SecretParts) == 0 {
+					t.Fatalf("no secret parts present: \n %+v", got[i])
+				}
 				if (got[i].VerificationError() != nil) != tt.wantVerificationErr {
 					t.Fatalf("wantVerificationError = %v, verification error = %v", tt.wantVerificationErr, got[i].VerificationError())
 				}
 			}
-			ignoreOpts := cmpopts.IgnoreFields(detectors.Result{}, "Raw", "verificationError")
+			ignoreOpts := cmpopts.IgnoreFields(detectors.Result{}, "Raw", "verificationError", "primarySecret", "SecretParts", "chunkOffset", "chunkOffsetSet")
 			if diff := cmp.Diff(got, tt.want, ignoreOpts); diff != "" {
 				t.Errorf("Weightsandbiases.FromData() %s diff: (-got +want)\n%s", tt.name, diff)
 			}

pkg/detectors/weightsandbiases/v2/weightsandbiases.go

@@ -0,0 +1,31 @@
+package weightsandbiases
+
+import (
+	"context"
+
+	regexp "github.com/wasilibs/go-re2"
+
+	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors"
+	base "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/weightsandbiases"
+)
+
+type Scanner struct {
+	base.BaseScanner
+}
+
+// Ensure the Scanner satisfies the interface at compile time.
+var _ detectors.Detector = (*Scanner)(nil)
+var _ detectors.Versioner = (*Scanner)(nil)
+
+// wandb_v1_<27 alphanumeric chars>_<49 alphanumeric chars>
+// Example: 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
+var keyPat = regexp.MustCompile(`\b(wandb_v1_[A-Za-z0-9]{27}_[A-Za-z0-9]{49})\b`)
+
+func (s Scanner) Version() int { return 2 }
+
+func (s Scanner) Keywords() []string { return []string{"wandb_v1_"} }
+
+func (s Scanner) FromData(ctx context.Context, verify bool, data []byte) ([]detectors.Result, error) {
+	return s.BaseScanner.FromData(ctx, verify, data, keyPat, s.Version())
+}
+

pkg/detectors/weightsandbiases/v2/weightsandbiases_integration_test.go

@@ -0,0 +1,175 @@
+//go:build detectors
+// +build detectors
+
+package weightsandbiases
+
+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"
+	base "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/weightsandbiases"
+	"github.com/trufflesecurity/trufflehog/v3/pkg/pb/detector_typepb"
+)
+
+func TestWeightsandbiasesV2_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)
+	}
+	secret := testSecrets.MustGetField("WEIGHTSANDBIASES_V2")
+	inactiveSecret := "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"
+
+	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:   []byte(fmt.Sprintf("You can find a weightsandbiases secret %s within", secret)),
+				verify: true,
+			},
+			want: []detectors.Result{
+				{
+					DetectorType: detector_typepb.DetectorType_WeightsAndBiases,
+					Verified:     true,
+					ExtraData: map[string]string{
+						"admin":    "false",
+						"email":    "[email protected]",
+						"username": "muneeb-khan-222",
+						"version":  "2",
+					},
+				},
+			},
+			wantErr:             false,
+			wantVerificationErr: false,
+		},
+		{
+			name: "found, unverified",
+			s:    Scanner{},
+			args: args{
+				ctx:    context.Background(),
+				data:   []byte(fmt.Sprintf("You can find a weightsandbiases secret %s within but not valid", inactiveSecret)),
+				verify: true,
+			},
+			want: []detectors.Result{
+				{
+					DetectorType: detector_typepb.DetectorType_WeightsAndBiases,
+					Verified:     false,
+					ExtraData:    map[string]string{"version": "2"},
+				},
+			},
+			wantErr:             false,
+			wantVerificationErr: 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,
+			wantVerificationErr: false,
+		},
+		{
+			name: "found, would be verified if not for timeout",
+			s:    Scanner{BaseScanner: base.BaseScanner{Client: common.SaneHttpClientTimeOut(1 * time.Microsecond)}},
+			args: args{
+				ctx:    context.Background(),
+				data:   []byte(fmt.Sprintf("You can find a weightsandbiases secret %s within", secret)),
+				verify: true,
+			},
+			want: []detectors.Result{
+				{
+					DetectorType: detector_typepb.DetectorType_WeightsAndBiases,
+					Verified:     false,
+					ExtraData:    map[string]string{"version": "2"},
+				},
+			},
+			wantErr:             false,
+			wantVerificationErr: true,
+		},
+		{
+			name: "found, verified but unexpected api surface",
+			s:    Scanner{BaseScanner: base.BaseScanner{Client: common.ConstantResponseHttpClient(404, "")}},
+			args: args{
+				ctx:    context.Background(),
+				data:   []byte(fmt.Sprintf("You can find a weightsandbiases secret %s within", secret)),
+				verify: true,
+			},
+			want: []detectors.Result{
+				{
+					DetectorType: detector_typepb.DetectorType_WeightsAndBiases,
+					Verified:     false,
+					ExtraData:    map[string]string{"version": "2"},
+				},
+			},
+			wantErr:             false,
+			wantVerificationErr: true,
+		},
+	}
+
+	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("WeightsandbiasesV2.FromData() error = %v, wantErr %v", err, tt.wantErr)
+				return
+			}
+			for i := range got {
+				if len(got[i].Raw) == 0 {
+					t.Fatalf("no raw secret present: \n %+v", got[i])
+				}
+				if len(got[i].SecretParts) == 0 {
+					t.Fatalf("no secret parts present: \n %+v", got[i])
+				}
+				if (got[i].VerificationError() != nil) != tt.wantVerificationErr {
+					t.Fatalf("wantVerificationError = %v, verification error = %v", tt.wantVerificationErr, got[i].VerificationError())
+				}
+			}
+			ignoreOpts := cmpopts.IgnoreFields(detectors.Result{}, "Raw", "verificationError", "primarySecret", "SecretParts", "chunkOffset", "chunkOffsetSet")
+			if diff := cmp.Diff(got, tt.want, ignoreOpts); diff != "" {
+				t.Errorf("WeightsandbiasesV2.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)
+				}
+			}
+		})
+	}
+}

pkg/detectors/weightsandbiases/v2/weightsandbiases_test.go

@@ -0,0 +1,88 @@
+package weightsandbiases
+
+import (
+	"context"
+	"testing"
+
+	"github.com/google/go-cmp/cmp"
+
+	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors"
+	"github.com/trufflesecurity/trufflehog/v3/pkg/engine/ahocorasick"
+)
+
+func TestWeightsandbiasesV2_Pattern(t *testing.T) {
+	d := Scanner{}
+	ahoCorasickCore := ahocorasick.NewAhoCorasickCore([]detectors.Detector{d})
+	tests := []struct {
+		name  string
+		input string
+		want  []string
+	}{
+		{
+			name:  "typical pattern",
+			input: "WANDB_API_KEY = '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'",
+			want:  []string{"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"},
+		},
+		{
+			name: "finds all matches",
+			input: `WANDB_API_KEY = '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'
+WANDB_API_KEY = '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'`,
+			want: []string{
+				"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",
+				"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",
+			},
+		},
+		{
+			name:  "invalid pattern - wrong length",
+			input: "WANDB_API_KEY = 'wandb_v1_CNskTdKUs0f1uHZ4eOECFLof6aC_4IlqrKmMuTTfwXd5n6hf8VvcOX67MNii'",
+			want:  []string{},
+		},
+		{
+			name:  "does not match v1 legacy format",
+			input: "WANDB_API_KEY = 'eedf1c984f6b995ec40ecc6658356044847ffb31'",
+			want:  []string{},
+		},
+	}
+
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			matchedDetectors := ahoCorasickCore.FindDetectorMatches([]byte(test.input))
+			if len(test.want) > 0 && len(matchedDetectors) == 0 {
+				t.Errorf("keywords '%v' not matched by: %s", d.Keywords(), test.input)
+				return
+			}
+
+			results, err := d.FromData(context.Background(), false, []byte(test.input))
+			if err != nil {
+				t.Errorf("error = %v", err)
+				return
+			}
+
+			if len(results) != len(test.want) {
+				if len(results) == 0 {
+					t.Errorf("did not receive result")
+				} else {
+					t.Errorf("expected %d results, only received %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)
+			}
+		})
+	}
+}

pkg/engine/defaults/defaults.go

@@ -864,7 +864,8 @@ import (
 	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors/webscraper"
 	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors/webscraping"
 	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors/websitepulse"
-	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors/weightsandbiases"
+	weightsandbiasesv1 "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/weightsandbiases/v1"
+	weightsandbiasesv2 "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/weightsandbiases/v2"
 	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors/whoxy"
 	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors/wistia"
 	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors/wit"
@@ -1789,7 +1790,8 @@ func buildDetectorList() []detectors.Detector {
 		&webscraper.Scanner{},
 		&webscraping.Scanner{},
 		&websitepulse.Scanner{},
-		&weightsandbiases.Scanner{},
+		&weightsandbiasesv1.Scanner{},
+		&weightsandbiasesv2.Scanner{},
 		// &wepay.Scanner{},
 		&whoxy.Scanner{},
 		&wistia.Scanner{},
@@ -1892,6 +1894,8 @@ func buildDetectorList() []detectors.Detector {
 			return !feature.MSTeamsWebhookV2DetectorEnabled.Load()
 		case *solarwindsobservability.Scanner:
 			return !feature.SolarwindsDetectorEnabled.Load()
+		case *weightsandbiasesv2.Scanner:
+			return !feature.WeightsAndBiasesV2DetectorEnabled.Load()
 		default:
 			return false
 		}

pkg/feature/feature.go

@@ -50,6 +50,7 @@ var (
 	NewRelicMobileAppTokenDetectorEnabled    atomic.Bool
 	MSTeamsWebhookV2DetectorEnabled          atomic.Bool
 	SolarwindsDetectorEnabled                atomic.Bool
+	WeightsAndBiasesV2DetectorEnabled        atomic.Bool
 )
 
 type AtomicString struct {