diff --git a/main.go b/main.go index 549b942e7..d55fae698 100644 --- a/main.go +++ b/main.go @@ -538,6 +538,7 @@ func run(state overseer.State, logSync func() error) { feature.WitDetectorEnabled.Store(true) feature.RevDetectorEnabled.Store(true) feature.UserDetectorEnabled.Store(true) + feature.BraintrustDetectorEnabled.Store(true) conf := &config.Config{} if *configFilename != "" { diff --git a/pkg/detectors/braintrust/braintrust.go b/pkg/detectors/braintrust/braintrust.go new file mode 100644 index 000000000..7293db4c2 --- /dev/null +++ b/pkg/detectors/braintrust/braintrust.go @@ -0,0 +1,135 @@ +package braintrust + +import ( + "context" + "fmt" + "io" + "net/http" + + 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 +} + +// Compile-time interface check +var _ detectors.Detector = (*Scanner)(nil) + +var ( + defaultClient = common.SaneHttpClient() + + // Braintrust API keys: + // Format: sk- + 48 alphanumeric characters (observed) + braintrustTokenPat = regexp.MustCompile( + `\b(sk-[A-Za-z0-9]{48})\b`, + ) +) + +// Keywords used for fast pre-filtering +func (s Scanner) Keywords() []string { + return []string{"braintrust"} +} + +func (s Scanner) getClient() *http.Client { + if s.client != nil { + return s.client + } + return defaultClient +} + +// FromData scans for Braintrust API tokens and optionally verifies them +func (s Scanner) FromData( + ctx context.Context, + verify bool, + data []byte, +) (results []detectors.Result, err error) { + + dataStr := string(data) + + uniqueTokens := make(map[string]struct{}) + for _, match := range braintrustTokenPat.FindAllStringSubmatch(dataStr, -1) { + uniqueTokens[match[1]] = struct{}{} + } + + for token := range uniqueTokens { + result := detectors.Result{ + DetectorType: detector_typepb.DetectorType_BrainTrustApiKey, + Raw: []byte(token), + Redacted: token[:8] + "...", + SecretParts: map[string]string{ + "token": token, + }, + } + + if verify { + verified, verificationErr := verifyBraintrustToken( + ctx, + s.getClient(), + token, + ) + result.SetVerificationError(verificationErr, token) + result.Verified = verified + } + + results = append(results, result) + } + + return +} + +func verifyBraintrustToken( + ctx context.Context, + client *http.Client, + token string, +) (bool, error) { + + req, err := http.NewRequestWithContext( + ctx, + http.MethodGet, + "https://api.braintrust.dev/v1/project?limit=1", + http.NoBody, + ) + if err != nil { + return false, err + } + + req.Header.Set("Authorization", "Bearer "+token) + + 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 or revoked token + return false, nil + case http.StatusForbidden: + // Valid token but lacks permission (still valid) + return true, nil + default: + return false, fmt.Errorf( + "unexpected HTTP response status %d", + res.StatusCode, + ) + } +} + +func (s Scanner) Type() detector_typepb.DetectorType { + return detector_typepb.DetectorType_BrainTrustApiKey +} + +func (s Scanner) Description() string { + return "Braintrust is an AI evaluation and observability platform. Braintrust API keys can be used to access evaluation, observability, and AI development resources within the Braintrust platform." +} diff --git a/pkg/detectors/braintrust/braintrust_integration_test.go b/pkg/detectors/braintrust/braintrust_integration_test.go new file mode 100644 index 000000000..72f7e288f --- /dev/null +++ b/pkg/detectors/braintrust/braintrust_integration_test.go @@ -0,0 +1,181 @@ +//go:build detectors +// +build detectors + +package braintrust + +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 TestBraintrust_FromData(t *testing.T) { + ctx, cancel := context.WithTimeout(context.Background(), time.Second*10) + defer cancel() + + // Load secrets from GCP (same pattern as other detectors) + testSecrets, err := common.GetSecret(ctx, "trufflehog-testing", "detectors6") + if err != nil { + t.Fatalf("could not get test secrets from GCP: %s", err) + } + + activeToken := testSecrets.MustGetField("BRAINTRUST_API_KEY") + inactiveToken := "sk-AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA" + + 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{}, "Using Braintrust API key %s", activeToken), + verify: true, + }, + want: []detectors.Result{ + { + DetectorType: detector_typepb.DetectorType_BrainTrustApiKey, + Verified: true, + Raw: []byte(activeToken), + Redacted: activeToken[:8] + "...", + }, + }, + }, + { + name: "found, real token, verification error due to timeout", + s: Scanner{client: common.SaneHttpClientTimeOut(1 * time.Microsecond)}, + args: args{ + ctx: context.Background(), + data: fmt.Appendf([]byte{}, "Using Braintrust API key %s", activeToken), + verify: true, + }, + want: []detectors.Result{ + { + DetectorType: detector_typepb.DetectorType_BrainTrustApiKey, + Verified: false, + Raw: []byte(activeToken), + Redacted: activeToken[:8] + "...", + }, + }, + wantVerificationErr: true, + }, + { + name: "found, real token, verification error due to unexpected api surface", + s: Scanner{client: common.ConstantResponseHttpClient(500, "{}")}, + args: args{ + ctx: context.Background(), + data: fmt.Appendf([]byte{}, "Using Braintrust API key %s", activeToken), + verify: true, + }, + want: []detectors.Result{ + { + DetectorType: detector_typepb.DetectorType_BrainTrustApiKey, + Verified: false, + Raw: []byte(activeToken), + Redacted: activeToken[:8] + "...", + }, + }, + wantVerificationErr: true, + }, + { + name: "found, unverified (inactive token)", + s: Scanner{}, + args: args{ + ctx: context.Background(), + data: fmt.Appendf([]byte{}, "Using Braintrust API key %s", inactiveToken), + verify: true, + }, + want: []detectors.Result{ + { + DetectorType: detector_typepb.DetectorType_BrainTrustApiKey, + Verified: false, + Raw: []byte(inactiveToken), + Redacted: inactiveToken[:8] + "...", + }, + }, + }, + { + name: "not found", + s: Scanner{}, + args: args{ + ctx: context.Background(), + data: []byte("no secrets here"), + verify: true, + }, + want: nil, + }, + } + + 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.Fatalf("Braintrust.FromData() error = %v, wantErr %v", err, tt.wantErr) + } + + 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", + "chunkOffset", + "chunkOffsetSet", + ) + + if diff := cmp.Diff(got, tt.want, ignoreOpts); diff != "" { + t.Errorf("Braintrust.FromData() %s diff: (-got +want)\n%s", tt.name, diff) + } + }) + } +} + +func BenchmarkBraintrust_FromData(b *testing.B) { + ctx := context.Background() + s := Scanner{} + + for name, data := range detectors.MustGetBenchmarkData() { + b.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/braintrust/braintrust_test.go b/pkg/detectors/braintrust/braintrust_test.go new file mode 100644 index 000000000..37f26b6db --- /dev/null +++ b/pkg/detectors/braintrust/braintrust_test.go @@ -0,0 +1,136 @@ +package braintrust + +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 TestBraintrust_Pattern(t *testing.T) { + d := Scanner{} + ahoCorasickCore := ahocorasick.NewAhoCorasickCore([]detectors.Detector{d}) + + tests := []struct { + name string + input string + want []string + }{ + { + name: "valid pattern - basic", + input: ` + [INFO] Starting AI eval + [DEBUG] braintrust_key=sk-JMwsKdC5JJfEGS4NhbQjzh1zfeYvGCuBkosFRWn98Z1H13Yg + `, + want: []string{ + "sk-JMwsKdC5JJfEGS4NhbQjzh1zfeYvGCuBkosFRWn98Z1H13Yg", + }, + }, + { + name: "valid pattern - with keyword nearby", + input: ` + [INFO] braintrust initialized + [DEBUG] BRAINTRUST_API_KEY=sk-76cnJ2Ns8wHZao70KdUdlZpBuSzej8gEokToNyeSPtd1RyZB + `, + want: []string{ + "sk-76cnJ2Ns8wHZao70KdUdlZpBuSzej8gEokToNyeSPtd1RyZB", + }, + }, + { + name: "valid pattern - multiple tokens", + input: ` + braintrust_key1=sk-AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA + braintrust_key2=sk-BBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBB + `, + want: []string{ + "sk-AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA", + "sk-BBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBB", + }, + }, + { + name: "invalid pattern - too short", + input: ` + [DEBUG] braintrust_key=sk-1234 + `, + want: nil, + }, + { + name: "invalid pattern - too long", + input: ` + [DEBUG] braintrust_key=sk-AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA + `, + want: nil, + }, + { + name: "invalid pattern - invalid characters", + input: ` + [DEBUG] braintrust_key=sk-AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA! + `, + want: nil, + }, + { + name: "invalid pattern - keyword only", + input: ` + [INFO] braintrust API initialized sk- + `, + want: nil, + }, + { + name: "valid pattern - mixed alphanumeric", + input: ` + [DEBUG] braintrust_key=sk-a1B2c3D4e5F6g7H8i9J0kLmNoPqRsTuVwXyZ1234567890ao + `, + want: []string{ + "sk-a1B2c3D4e5F6g7H8i9J0kLmNoPqRsTuVwXyZ1234567890ao", + }, + }, + } + + 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 f3604c038..537d2332b 100644 --- a/pkg/engine/defaults/defaults.go +++ b/pkg/engine/defaults/defaults.go @@ -115,6 +115,7 @@ import ( "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/box" "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/boxoauth" "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/braintreepayments" + "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/braintrust" brandfetchv1 "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/brandfetch/v1" brandfetchv2 "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/brandfetch/v2" "github.com/trufflesecurity/trufflehog/v3/pkg/detectors/browserstack" @@ -997,6 +998,7 @@ func buildDetectorList() []detectors.Detector { &box.Scanner{}, &boxoauth.Scanner{}, &braintreepayments.Scanner{}, + &braintrust.Scanner{}, &brandfetchv1.Scanner{}, &brandfetchv2.Scanner{}, &browserstack.Scanner{}, @@ -1800,6 +1802,8 @@ func buildDetectorList() []detectors.Detector { return !feature.RevDetectorEnabled.Load() case *user.Scanner: return !feature.UserDetectorEnabled.Load() + case *braintrust.Scanner: + return !feature.BraintrustDetectorEnabled.Load() default: return false } diff --git a/pkg/engine/defaults/defaults_test.go b/pkg/engine/defaults/defaults_test.go index 1983e7b4f..faec737f9 100644 --- a/pkg/engine/defaults/defaults_test.go +++ b/pkg/engine/defaults/defaults_test.go @@ -125,15 +125,16 @@ var excludedFromDefaultList = map[detector_typepb.DetectorType]struct{}{ // Feature flag gated detectors // These should be removed from this list when we remove the feature flag - detector_typepb.DetectorType_Cloudinary: {}, - detector_typepb.DetectorType_DatadogApikey: {}, - detector_typepb.DetectorType_Enigma: {}, - detector_typepb.DetectorType_GitLabOauth2: {}, - detector_typepb.DetectorType_Pinecone: {}, - detector_typepb.DetectorType_TLy: {}, - detector_typepb.DetectorType_Wit: {}, - detector_typepb.DetectorType_Rev: {}, - detector_typepb.DetectorType_User: {}, + detector_typepb.DetectorType_Cloudinary: {}, + detector_typepb.DetectorType_DatadogApikey: {}, + detector_typepb.DetectorType_Enigma: {}, + detector_typepb.DetectorType_GitLabOauth2: {}, + detector_typepb.DetectorType_Pinecone: {}, + detector_typepb.DetectorType_TLy: {}, + detector_typepb.DetectorType_Wit: {}, + detector_typepb.DetectorType_Rev: {}, + detector_typepb.DetectorType_User: {}, + detector_typepb.DetectorType_BrainTrustApiKey: {}, // Reserved / special types. detector_typepb.DetectorType_CustomRegex: {}, // added dynamically via engine config, not via buildDetectorList() diff --git a/pkg/feature/feature.go b/pkg/feature/feature.go index a44e0526d..c8585bb89 100644 --- a/pkg/feature/feature.go +++ b/pkg/feature/feature.go @@ -25,6 +25,7 @@ var ( WitDetectorEnabled atomic.Bool RevDetectorEnabled atomic.Bool UserDetectorEnabled atomic.Bool + BraintrustDetectorEnabled atomic.Bool ) type AtomicString struct { diff --git a/pkg/pb/detector_typepb/detector_type.pb.go b/pkg/pb/detector_typepb/detector_type.pb.go index 11428679d..1685583e8 100644 --- a/pkg/pb/detector_typepb/detector_type.pb.go +++ b/pkg/pb/detector_typepb/detector_type.pb.go @@ -1106,6 +1106,7 @@ const ( DetectorType_GitLabOauth2 DetectorType = 1050 DetectorType_SpectralOps DetectorType = 1051 DetectorType_AWSAppSync DetectorType = 1052 + DetectorType_BrainTrustApiKey DetectorType = 1053 ) // Enum value maps for DetectorType. @@ -2160,6 +2161,7 @@ var ( 1050: "GitLabOauth2", 1051: "SpectralOps", 1052: "AWSAppSync", + 1053: "BrainTrustApiKey", } DetectorType_value = map[string]int32{ "Alibaba": 0, @@ -3211,6 +3213,7 @@ var ( "GitLabOauth2": 1050, "SpectralOps": 1051, "AWSAppSync": 1052, + "BrainTrustApiKey": 1053, } ) @@ -3246,7 +3249,7 @@ 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, 0xe2, 0x88, 0x01, 0x0a, 0x0c, 0x44, 0x65, 0x74, 0x65, 0x63, 0x74, + 0x74, 0x79, 0x70, 0x65, 0x2a, 0xf9, 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, @@ -4340,12 +4343,13 @@ var file_detector_type_proto_rawDesc = []byte{ 0x65, 0x10, 0x99, 0x08, 0x12, 0x11, 0x0a, 0x0c, 0x47, 0x69, 0x74, 0x4c, 0x61, 0x62, 0x4f, 0x61, 0x75, 0x74, 0x68, 0x32, 0x10, 0x9a, 0x08, 0x12, 0x10, 0x0a, 0x0b, 0x53, 0x70, 0x65, 0x63, 0x74, 0x72, 0x61, 0x6c, 0x4f, 0x70, 0x73, 0x10, 0x9b, 0x08, 0x12, 0x0f, 0x0a, 0x0a, 0x41, 0x57, 0x53, - 0x41, 0x70, 0x70, 0x53, 0x79, 0x6e, 0x63, 0x10, 0x9c, 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, + 0x41, 0x70, 0x70, 0x53, 0x79, 0x6e, 0x63, 0x10, 0x9c, 0x08, 0x12, 0x15, 0x0a, 0x10, 0x42, 0x72, + 0x61, 0x69, 0x6e, 0x54, 0x72, 0x75, 0x73, 0x74, 0x41, 0x70, 0x69, 0x4b, 0x65, 0x79, 0x10, 0x9d, + 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 6423f9ede..b4215f453 100644 --- a/proto/detector_type.proto +++ b/proto/detector_type.proto @@ -1054,4 +1054,5 @@ enum DetectorType { GitLabOauth2 = 1050; SpectralOps = 1051; AWSAppSync = 1052; + BrainTrustApiKey = 1053; }