Files
SnapOtter/apps/api/src/routes/fetch-urls.ts
T
SnapOtter 4e64ee2779 fix(security): comprehensive security audit and hardening
Auth: login rate limit 30/min (was 500), global rate limit 1000/min (was
unlimited), password/username max lengths on all Zod schemas, session
invalidation on role change, API key legacy scan bounded to 100 keys.

SVG: hardened regex sanitizer with CDATA stripping, XML entity decoding,
set/animate/iframe/embed blocking, comprehensive data: URI blocking,
use element external href blocking. 11 attack payload fixtures added.

SSRF: fixed DNS rebinding TOCTOU by pinning resolved IPs via custom
HTTP/HTTPS agents. Added 6to4 and NAT64 to blocked IPv6 ranges.

Docker: capability dropping (cap_drop ALL + minimal cap_add), resource
limits (4g/8g mem, 512/1024 pids), healthcheck timeout, password
removed from startup banner, default password warning comments.

Network: CSP and HSTS applied in all environments (not just production),
stack traces removed from all error responses, internal paths stripped
from error details, per-route rate limits on uploads (60/min) and URL
fetches (200/hour).

Files: exclusive temp file creation (O_EXCL), disk space circuit
breaker, per-user storage quotas, settings payload 64KB size guard.

Python sidecar: script name allowlist in dispatcher, minimal environment
for subprocess spawns.

Dependencies: fixed 6 production CVEs (drizzle-orm, fastify, fast-uri,
@fastify/static, next, archiver/lodash). Pinned all GitHub Actions to
SHA hashes.

114 security tests added. Full OWASP Top 10 penetration test matrix
verified against production Docker container (30/30 pass after
hardening).
2026-05-13 21:33:50 +08:00

284 lines
8.0 KiB
TypeScript

/**
* Fetch URLs route.
*
* POST /api/v1/fetch-urls
*
* Accepts a JSON body with { urls: string[] } (1-50 URLs).
* Fetches each URL server-side with SSRF protection, validates as an image,
* saves to a workspace, generates a preview for non-browser formats, and
* returns results with download URLs.
*/
import { randomUUID } from "node:crypto";
import { writeFile } from "node:fs/promises";
import { basename, join } from "node:path";
import type { FastifyInstance } from "fastify";
import PQueue from "p-queue";
import sharp from "sharp";
import { z } from "zod";
import { validateImageBuffer } from "../lib/file-validation.js";
import { sanitizeFilename } from "../lib/filename.js";
import {
FETCH_TIMEOUT_MS,
MAX_URL_FETCH_SIZE,
MAX_URLS_PER_REQUEST,
safeFetch,
URL_FETCH_CONCURRENCY,
} from "../lib/ssrf.js";
import { createWorkspace } from "../lib/workspace.js";
/** Formats browsers can display natively (no preview needed). */
const BROWSER_PREVIEWABLE = new Set([
"image/jpeg",
"image/png",
"image/gif",
"image/webp",
"image/svg+xml",
"image/bmp",
"image/avif",
]);
/** Map detected format string to MIME type. */
const FORMAT_TO_MIME: Record<string, string> = {
jpeg: "image/jpeg",
png: "image/png",
gif: "image/gif",
webp: "image/webp",
svg: "image/svg+xml",
bmp: "image/bmp",
avif: "image/avif",
tiff: "image/tiff",
heif: "image/heic",
jxl: "image/jxl",
ico: "image/x-icon",
psd: "image/vnd.adobe.photoshop",
raw: "image/x-dcraw",
tga: "image/x-tga",
exr: "image/x-exr",
hdr: "image/vnd.radiance",
jp2: "image/jp2",
qoi: "image/x-qoi",
eps: "application/postscript",
dds: "image/x-dds",
cur: "image/x-icon",
dpx: "image/x-dpx",
fits: "image/fits",
ppm: "image/x-portable-pixmap",
pgm: "image/x-portable-graymap",
pbm: "image/x-portable-bitmap",
pfm: "image/x-portable-floatmap",
};
const fetchUrlsSchema = z.object({
urls: z
.array(z.string().url("Each entry must be a valid URL"))
.min(1, "At least one URL is required")
.max(MAX_URLS_PER_REQUEST, `Maximum ${MAX_URLS_PER_REQUEST} URLs per request`),
});
interface SuccessResult {
success: true;
url: string;
filename: string;
contentType: string;
size: number;
width: number;
height: number;
downloadUrl: string;
previewUrl: string | null;
}
interface FailureResult {
success: false;
url: string;
error: string;
}
type FetchResult = SuccessResult | FailureResult;
/**
* Extract a usable filename from a URL path, falling back to a UUID-based name.
*/
function filenameFromUrl(url: string): string {
try {
const pathname = new URL(url).pathname;
const base = basename(pathname);
// Decode percent-encoded characters
const decoded = decodeURIComponent(base);
// Only use it if it looks like a file with an extension
if (decoded?.includes(".") && decoded.length <= 255) {
return decoded;
}
} catch {
// ignore parse errors
}
return `image-${randomUUID().slice(0, 8)}`;
}
/**
* Return a filename that does not collide with any name already in `used`.
* Appends `_1`, `_2`, etc. before the extension when a collision is found.
* Mirrors the deduplication logic in batch.ts.
*/
function getUniqueName(name: string, used: Set<string>): string {
if (!used.has(name)) {
used.add(name);
return name;
}
const dotIdx = name.lastIndexOf(".");
const base = dotIdx > 0 ? name.slice(0, dotIdx) : name;
const ext = dotIdx > 0 ? name.slice(dotIdx) : "";
let counter = 1;
let candidate = `${base}_${counter}${ext}`;
while (used.has(candidate)) {
counter++;
candidate = `${base}_${counter}${ext}`;
}
used.add(candidate);
return candidate;
}
export async function registerFetchUrlsRoute(app: FastifyInstance): Promise<void> {
app.post(
"/api/v1/fetch-urls",
{ config: { rateLimit: { max: 200, timeWindow: "1 hour" } } },
async (request, reply) => {
// Validate body
const parsed = fetchUrlsSchema.safeParse(request.body);
if (!parsed.success) {
const messages = parsed.error.issues.map((i) => i.message).join("; ");
return reply.status(400).send({ error: messages });
}
const { urls } = parsed.data;
const jobId = randomUUID();
const workspace = await createWorkspace(jobId);
const outputDir = join(workspace, "output");
const queue = new PQueue({ concurrency: URL_FETCH_CONCURRENCY });
// Track filenames to prevent collisions when multiple URLs resolve to the
// same name (e.g. https://a.com/photo.jpg and https://b.com/photo.jpg).
const usedFilenames = new Set<string>();
// Pre-allocate result slots to preserve order
const resultSlots: FetchResult[] = new Array(urls.length);
await Promise.all(
urls.map((url, index) =>
queue.add(async () => {
resultSlots[index] = await fetchSingleUrl(url, jobId, outputDir, usedFilenames);
}),
),
);
return reply.send({ results: resultSlots });
},
);
}
async function fetchSingleUrl(
url: string,
jobId: string,
outputDir: string,
usedFilenames: Set<string>,
): Promise<FetchResult> {
try {
// Fetch with SSRF protection and timeout
const controller = new AbortController();
const timeout = setTimeout(() => controller.abort(), FETCH_TIMEOUT_MS);
let response: Response;
try {
response = await safeFetch(url, controller.signal);
} finally {
clearTimeout(timeout);
}
if (!response.ok) {
return {
success: false,
url,
error: `HTTP ${response.status} ${response.statusText}`,
};
}
// Read body with size limit
const chunks: Uint8Array[] = [];
let totalSize = 0;
if (!response.body) {
return { success: false, url, error: "Empty response body" };
}
const reader = response.body.getReader();
try {
while (true) {
const { done, value } = await reader.read();
if (done) break;
totalSize += value.byteLength;
if (totalSize > MAX_URL_FETCH_SIZE) {
reader.cancel();
return {
success: false,
url,
error: `File exceeds maximum size of ${MAX_URL_FETCH_SIZE / (1024 * 1024)}MB`,
};
}
chunks.push(value);
}
} finally {
reader.releaseLock();
}
const buffer = Buffer.concat(chunks);
if (buffer.length === 0) {
return { success: false, url, error: "Empty response body" };
}
// Derive filename from URL, deduplicating to prevent overwrites when
// multiple URLs resolve to the same name (all URLs share one workspace).
const rawFilename = filenameFromUrl(url);
const filename = getUniqueName(sanitizeFilename(rawFilename), usedFilenames);
// Validate as an image
const validation = await validateImageBuffer(buffer, filename);
if (!validation.valid) {
return { success: false, url, error: validation.reason };
}
// Save to output directory
await writeFile(join(outputDir, filename), buffer);
const contentType = FORMAT_TO_MIME[validation.format] ?? "application/octet-stream";
const downloadUrl = `/api/v1/download/${jobId}/${encodeURIComponent(filename)}`;
// Generate preview for non-browser formats
let previewUrl: string | null = null;
if (!BROWSER_PREVIEWABLE.has(contentType)) {
try {
const previewBuffer = await sharp(buffer).webp({ quality: 80 }).toBuffer();
const previewFilename = `preview-${filename.replace(/\.[^.]+$/, "")}.webp`;
await writeFile(join(outputDir, previewFilename), previewBuffer);
previewUrl = `/api/v1/download/${jobId}/${encodeURIComponent(previewFilename)}`;
} catch {
// Preview generation failed -- non-fatal, skip preview
}
}
return {
success: true,
url,
filename,
contentType,
size: buffer.length,
width: validation.width,
height: validation.height,
downloadUrl,
previewUrl,
};
} catch (err) {
const message = err instanceof Error ? err.message : "Unknown error";
return { success: false, url, error: message };
}
}