[INS-497] Add Pganalyze Read Key Detector (#4993)

* Add pganalyze read key detector * regen protos and ran gofmt
2026-06-29 11:36:40 +05:00
parent c1a1d6a28a
commit c09d72666a
9 changed files with 537 additions and 27 deletions
@@ -543,6 +543,7 @@ func run(state overseer.State, logSync func() error) {
 	feature.RevDetectorEnabled.Store(true)
 	feature.UserDetectorEnabled.Store(true)
 	feature.BraintrustDetectorEnabled.Store(true)
+	feature.PgAnalyzeReadKeyDetectorEnabled.Store(true)

 	conf := &config.Config{}
 	if *configFilename != "" {
@@ -0,0 +1,144 @@
+package pganalyzereadkey
+
+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()
+
+	// pganalyze Read API keys use the format:
+	// pgar_<27 alphanumeric characters>
+	//
+	// Example:
+	// pgar_abcdefghijklmnopqrstuvwxyz12
+	pganalyzeTokenPat = regexp.MustCompile(
+		`\b(pgar_[A-Za-z0-9]{27})\b`,
+	)
+)
+
+// Keywords used for fast pre-filtering
+func (s Scanner) Keywords() []string {
+	return []string{
+		"pgar_",
+	}
+}
+
+func (s Scanner) getClient() *http.Client {
+	if s.client != nil {
+		return s.client
+	}
+	return defaultClient
+}
+
+// FromData scans for pganalyze 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{})
+
+	matches := pganalyzeTokenPat.FindAllStringSubmatch(dataStr, -1)
+	for _, match := range matches {
+		uniqueTokens[match[1]] = struct{}{}
+	}
+
+	for token := range uniqueTokens {
+		result := detectors.Result{
+			DetectorType: detector_typepb.DetectorType_PgAnalyzeReadKey,
+			Raw:          []byte(token),
+			SecretParts: map[string]string{
+				"key":         token,
+				"access_type": "read",
+			},
+		}
+
+		if verify {
+			verified, verificationErr := verifyPganalyzeToken(
+				ctx,
+				s.getClient(),
+				token,
+			)
+
+			result.SetVerificationError(verificationErr, token)
+			result.Verified = verified
+		}
+
+		results = append(results, result)
+	}
+
+	return
+}
+
+func verifyPganalyzeToken(
+	ctx context.Context,
+	client *http.Client,
+	token string,
+) (bool, error) {
+
+	req, err := http.NewRequestWithContext(
+		ctx,
+		http.MethodPost,
+		"https://app.pganalyze.com/graphql",
+		http.NoBody,
+	)
+	if err != nil {
+		return false, err
+	}
+
+	req.Header.Set("Authorization", "Token "+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
+
+	// Explicit invalid auth
+	case http.StatusUnauthorized:
+		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_PgAnalyzeReadKey
+}
+
+func (s Scanner) Description() string {
+	return "pganalyze is a PostgreSQL monitoring and performance analysis platform. pganalyze Read API keys can be used to access read-only monitoring and query performance data."
+}
@@ -0,0 +1,215 @@
+//go:build detectors
+// +build detectors
+
+package pganalyzereadkey
+
+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 TestPgAnalyzeReadKey_FromData(t *testing.T) {
+	ctx, cancel := context.WithTimeout(context.Background(), time.Second*10)
+	defer cancel()
+
+	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("PGANALYZE_READ_KEY")
+	inactiveToken := "pgar_abcdefghijklmnopqrstuvwxyz1"
+
+	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 pganalyze read key %s for dashboard access",
+					activeToken,
+				),
+				verify: true,
+			},
+			want: []detectors.Result{
+				{
+					DetectorType: detector_typepb.DetectorType_PgAnalyzeReadKey,
+					Verified:     true,
+					Raw:          []byte(activeToken),
+				},
+			},
+		},
+		{
+			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 pganalyze read key %s for dashboard access",
+					activeToken,
+				),
+				verify: true,
+			},
+			want: []detectors.Result{
+				{
+					DetectorType: detector_typepb.DetectorType_PgAnalyzeReadKey,
+					Verified:     false,
+					Raw:          []byte(activeToken),
+				},
+			},
+			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 pganalyze read key %s for dashboard access",
+					activeToken,
+				),
+				verify: true,
+			},
+			want: []detectors.Result{
+				{
+					DetectorType: detector_typepb.DetectorType_PgAnalyzeReadKey,
+					Verified:     false,
+					Raw:          []byte(activeToken),
+				},
+			},
+			wantVerificationErr: true,
+		},
+		{
+			name: "found, unverified (inactive token)",
+			s:    Scanner{},
+			args: args{
+				ctx: context.Background(),
+				data: fmt.Appendf(
+					[]byte{},
+					"Using pganalyze read key %s for dashboard access",
+					inactiveToken,
+				),
+				verify: true,
+			},
+			want: []detectors.Result{
+				{
+					DetectorType: detector_typepb.DetectorType_PgAnalyzeReadKey,
+					Verified:     false,
+					Raw:          []byte(inactiveToken),
+				},
+			},
+		},
+		{
+			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(
+					"PgAnalyzeReadKey.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(
+					"PgAnalyzeReadKey.FromData() %s diff: (-got +want)\n%s",
+					tt.name,
+					diff,
+				)
+			}
+		})
+	}
+}
+
+func BenchmarkPgAnalyzeReadKey_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)
+				}
+			}
+		})
+	}
+}
@@ -0,0 +1,138 @@
+package pganalyzereadkey
+
+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 TestPgAnalyzeReadKey_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] pganalyze initialized
+				[DEBUG] token=pgar_abcdefghijklmnopqrstuvwxyz1
+				[INFO] done
+			`,
+			want: []string{
+				"pgar_abcdefghijklmnopqrstuvwxyz1",
+			},
+		},
+		{
+			name: "valid pattern - env variable",
+			input: `
+				PGANALYZE_READ_KEY=pgar_123456789012345678901234567
+			`,
+			want: []string{
+				"pgar_123456789012345678901234567",
+			},
+		},
+		{
+			name: "valid pattern - multiple tokens",
+			input: `
+				pgar_aaaaaaaaaaaaaaaaaaaaaaaaaaa
+				pgar_bbbbbbbbbbbbbbbbbbbbbbbbbbb
+			`,
+			want: []string{
+				"pgar_aaaaaaaaaaaaaaaaaaaaaaaaaaa",
+				"pgar_bbbbbbbbbbbbbbbbbbbbbbbbbbb",
+			},
+		},
+		{
+			name: "invalid pattern - uppercase prefix",
+			input: `
+				PGAR_abcdefghijklmnopqrstuvwxyz1
+			`,
+			want: nil,
+		},
+		{
+			name: "invalid pattern - too short",
+			input: `
+				pgar_12345
+			`,
+			want: nil,
+		},
+		{
+			name: "invalid pattern - too long",
+			input: `
+				pgar_abcdefghijklmnopqrstuvwxyz123456
+			`,
+			want: nil,
+		},
+		{
+			name: "invalid pattern - invalid characters",
+			input: `
+				pgar_abcde!ghijklmnopqrstuvwxyz12
+			`,
+			want: nil,
+		},
+		{
+			name: "invalid pattern - keyword only",
+			input: `
+				pgar_
+			`,
+			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)
+			}
+		})
+	}
+}
@@ -563,6 +563,7 @@ import (
 	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors/peopledatalabs"
 	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors/pepipost"
 	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors/percy"
+	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors/pganalyzereadkey"
 	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors/photoroom"
 	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors/phraseaccesstoken"
 	"github.com/trufflesecurity/trufflehog/v3/pkg/detectors/pinata"
@@ -1461,6 +1462,7 @@ func buildDetectorList() []detectors.Detector {
 		&peopledatalabs.Scanner{},
 		&pepipost.Scanner{},
 		&percy.Scanner{},
+		&pganalyzereadkey.Scanner{},
 		&photoroom.Scanner{},
 		&phraseaccesstoken.Scanner{},
 		&pinata.Scanner{},
@@ -1804,6 +1806,8 @@ func buildDetectorList() []detectors.Detector {
 			return !feature.UserDetectorEnabled.Load()
 		case *braintrust.Scanner:
 			return !feature.BraintrustDetectorEnabled.Load()
+		case *pganalyzereadkey.Scanner:
+			return !feature.PgAnalyzeReadKeyDetectorEnabled.Load()
 		default:
 			return false
 		}
@@ -135,6 +135,7 @@ var excludedFromDefaultList = map[detector_typepb.DetectorType]struct{}{
 	detector_typepb.DetectorType_Rev:              {},
 	detector_typepb.DetectorType_User:             {},
 	detector_typepb.DetectorType_BrainTrustApiKey: {},
+	detector_typepb.DetectorType_PgAnalyzeReadKey: {},

 	// Reserved / special types.
 	detector_typepb.DetectorType_CustomRegex: {}, // added dynamically via engine config, not via buildDetectorList()
@@ -26,6 +26,7 @@ var (
 	RevDetectorEnabled              atomic.Bool
 	UserDetectorEnabled             atomic.Bool
 	BraintrustDetectorEnabled       atomic.Bool
+	PgAnalyzeReadKeyDetectorEnabled atomic.Bool
 )

 type AtomicString struct {
@@ -1107,6 +1107,7 @@ const (
 	DetectorType_SpectralOps                             DetectorType = 1051
 	DetectorType_AWSAppSync                              DetectorType = 1052
 	DetectorType_BrainTrustApiKey                        DetectorType = 1053
+	DetectorType_PgAnalyzeReadKey                        DetectorType = 1054
 )

 // Enum value maps for DetectorType.
@@ -2162,6 +2163,7 @@ var (
 		1051: "SpectralOps",
 		1052: "AWSAppSync",
 		1053: "BrainTrustApiKey",
+		1054: "PgAnalyzeReadKey",
 	}
 	DetectorType_value = map[string]int32{
 		"Alibaba":                               0,
@@ -3214,6 +3216,7 @@ var (
 		"SpectralOps":                       1051,
 		"AWSAppSync":                        1052,
 		"BrainTrustApiKey":                  1053,
+		"PgAnalyzeReadKey":                  1054,
 	}
 )

@@ -3249,7 +3252,7 @@ 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, 0xf9, 0x88, 0x01, 0x0a, 0x0c, 0x44, 0x65, 0x74, 0x65, 0x63, 0x74,
+	0x74, 0x79, 0x70, 0x65, 0x2a, 0x90, 0x89, 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,
@@ -4345,11 +4348,13 @@ var file_detector_type_proto_rawDesc = []byte{
 	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, 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,
+	0x08, 0x12, 0x15, 0x0a, 0x10, 0x50, 0x67, 0x41, 0x6e, 0x61, 0x6c, 0x79, 0x7a, 0x65, 0x52, 0x65,
+	0x61, 0x64, 0x4b, 0x65, 0x79, 0x10, 0x9e, 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 (
@@ -1055,4 +1055,5 @@ enum DetectorType {
  SpectralOps = 1051;
  AWSAppSync = 1052;
  BrainTrustApiKey = 1053;
+  PgAnalyzeReadKey = 1054;
 }