Add size limits on request bodies and upstream metadata reads

internal/handler/composer.go

@@ -108,7 +108,7 @@ func (h *ComposerHandler) handlePackageMetadata(w http.ResponseWriter, r *http.R
 		return
 	}
 
-	body, err := io.ReadAll(resp.Body)
+	body, err := ReadMetadata(resp.Body)
 	if err != nil {
 		http.Error(w, "failed to read response", http.StatusInternalServerError)
 		return

internal/handler/handler.go

@@ -30,6 +30,17 @@ func containsPathTraversal(path string) bool {
 	return false
 }
 
+// maxMetadataSize is the maximum size of upstream metadata responses (50 MB).
+// Package metadata (e.g. npm with many versions) can be large, but unbounded
+// reads risk OOM if an upstream misbehaves.
+const maxMetadataSize = 50 << 20
+
+// ReadMetadata reads an upstream response body with a size limit to prevent OOM
+// from unexpectedly large responses.
+func ReadMetadata(r io.Reader) ([]byte, error) {
+	return io.ReadAll(io.LimitReader(r, maxMetadataSize))
+}
+
 // Proxy provides shared functionality for protocol handlers.
 type Proxy struct {
 	DB       *database.DB

internal/handler/npm.go

@@ -3,7 +3,6 @@ package handler
 import (
 	"encoding/json"
 	"fmt"
-	"io"
 	"net/http"
 	"net/url"
 	"sort"
@@ -93,7 +92,7 @@ func (h *NPMHandler) handlePackageMetadata(w http.ResponseWriter, r *http.Reques
 	}
 
 	// Parse and rewrite tarball URLs
-	body, err := io.ReadAll(resp.Body)
+	body, err := ReadMetadata(resp.Body)
 	if err != nil {
 		JSONError(w, http.StatusInternalServerError, "failed to read response")
 		return

internal/handler/nuget.go

@@ -77,7 +77,7 @@ func (h *NuGetHandler) handleServiceIndex(w http.ResponseWriter, r *http.Request
 		return
 	}
 
-	body, err := io.ReadAll(resp.Body)
+	body, err := ReadMetadata(resp.Body)
 	if err != nil {
 		http.Error(w, "failed to read response", http.StatusInternalServerError)
 		return

internal/handler/pub.go

@@ -109,7 +109,7 @@ func (h *PubHandler) handlePackageMetadata(w http.ResponseWriter, r *http.Reques
 		return
 	}
 
-	body, err := io.ReadAll(resp.Body)
+	body, err := ReadMetadata(resp.Body)
 	if err != nil {
 		http.Error(w, "failed to read response", http.StatusInternalServerError)
 		return

internal/handler/pypi.go

@@ -92,7 +92,7 @@ func (h *PyPIHandler) handleSimplePackage(w http.ResponseWriter, r *http.Request
 		return
 	}
 
-	body, err := io.ReadAll(resp.Body)
+	body, err := ReadMetadata(resp.Body)
 	if err != nil {
 		http.Error(w, "failed to read response", http.StatusInternalServerError)
 		return
@@ -259,7 +259,7 @@ func (h *PyPIHandler) proxyAndRewriteJSON(w http.ResponseWriter, r *http.Request
 		return
 	}
 
-	body, err := io.ReadAll(resp.Body)
+	body, err := ReadMetadata(resp.Body)
 	if err != nil {
 		http.Error(w, "failed to read response", http.StatusInternalServerError)
 		return

internal/handler/read_metadata_test.go

@@ -0,0 +1,34 @@
+package handler
+
+import (
+	"bytes"
+	"testing"
+)
+
+func TestReadMetadata(t *testing.T) {
+	t.Run("small body", func(t *testing.T) {
+		data := []byte("hello world")
+		got, err := ReadMetadata(bytes.NewReader(data))
+		if err != nil {
+			t.Fatalf("unexpected error: %v", err)
+		}
+		if !bytes.Equal(got, data) {
+			t.Errorf("got %q, want %q", got, data)
+		}
+	})
+
+	t.Run("truncates at limit", func(t *testing.T) {
+		// Create a reader slightly larger than maxMetadataSize
+		data := make([]byte, maxMetadataSize+100)
+		for i := range data {
+			data[i] = 'x'
+		}
+		got, err := ReadMetadata(bytes.NewReader(data))
+		if err != nil {
+			t.Fatalf("unexpected error: %v", err)
+		}
+		if len(got) != int(maxMetadataSize) {
+			t.Errorf("got length %d, want %d", len(got), maxMetadataSize)
+		}
+	})
+}

internal/server/api.go

@@ -327,6 +327,7 @@ func (h *APIHandler) HandleGetVulns(w http.ResponseWriter, r *http.Request) {
 // @Failure 500 {string} string
 // @Router /api/outdated [post]
 func (h *APIHandler) HandleOutdated(w http.ResponseWriter, r *http.Request) {
+	r.Body = http.MaxBytesReader(w, r.Body, 1<<20) // 1 MB
 	var req OutdatedRequest
 	if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
 		http.Error(w, "invalid request body", http.StatusBadRequest)
@@ -372,6 +373,7 @@ func (h *APIHandler) HandleOutdated(w http.ResponseWriter, r *http.Request) {
 // @Failure 500 {string} string
 // @Router /api/bulk [post]
 func (h *APIHandler) HandleBulkLookup(w http.ResponseWriter, r *http.Request) {
+	r.Body = http.MaxBytesReader(w, r.Body, 1<<20) // 1 MB
 	var req BulkRequest
 	if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
 		http.Error(w, "invalid request body", http.StatusBadRequest)

internal/server/api_test.go

@@ -94,6 +94,25 @@ func TestHandleOutdated_EmptyBody(t *testing.T) {
 	}
 }
 
+func TestHandleOutdated_OversizedBody(t *testing.T) {
+	logger := slog.New(slog.NewTextHandler(os.Stdout, nil))
+	svc := enrichment.New(logger)
+	h := NewAPIHandler(svc, nil)
+
+	// Send a body larger than 1 MB
+	body := make([]byte, 2<<20)
+	for i := range body {
+		body[i] = 'x'
+	}
+	req := httptest.NewRequest("POST", "/api/outdated", bytes.NewReader(body))
+	w := httptest.NewRecorder()
+	h.HandleOutdated(w, req)
+
+	if w.Code != http.StatusBadRequest {
+		t.Errorf("expected status %d for oversized body, got %d", http.StatusBadRequest, w.Code)
+	}
+}
+
 func TestHandleOutdated_InvalidJSON(t *testing.T) {
 	logger := slog.New(slog.NewTextHandler(os.Stdout, nil))
 	svc := enrichment.New(logger)