CVE-2026-64650 (GCVE-0-2026-64650)

Vulnerability from cvelistv5 – Published: 2026-07-20 20:27 – Updated: 2026-07-20 21:45
VLAI
Title
AI SDK Codex Harness Tool Relay Authorization Bypass
Summary
The `@ai-sdk/harness-opencode` tool is an HarnessV1 adapter backed by @openai/codex-sdk, which drives the codex command line interface. Prior to version 1.0.29, the tool relay authorizes requests from any process whose command line contains an allowed helper script path (the Codex CLI shim). This allows untrusted code executing in the sandbox to invoke arbitrary host-exposed tools, including secret lookups, deployment operations, and cloud API calls without a corresponding model-authorized tool-call event. Exploitation requires a Linux environment (the vulnerable fallback checks `process.platform === 'linux'` and reads `/proc`); an active harness session with one or more host-provided tools; and untrusted code executing in the sandbox (e.g. a malicious dependency, build script, or lifecycle hook) The fix in version 1.0.29 removes the process-path authorization fallback entirely. Relay requests are now only accepted after exact, short-lived, one-time authorization matching the tool name and input from a bridge-observed model event. Some workarounds are available. Do not run the Codex harness on untrusted repositories or with untrusted dependencies, and/or limit host-exposed tools to non-sensitive operations when working with untrusted code.
Severity
6.3 (Medium)
CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:N/VC:N/VI:L/VA:N/SC:H/SI:H/SA:N
SSVC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-07-20 21:45 UTC
CWE
  • CWE-863 - Incorrect Authorization
Assigner
GitHub_M CNA under the mitre root
CNA scorecard C 72/100 over 7838 records in the last 180 days details
References
Impacted products
Vendor Product Version CPE status
vercel @ai-sdk/harness-codex Affected: < 1.0.29
guessed Create a notification for this product.
Show details on NVD website
To clipboard 

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-64650",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-07-20T21:45:38.151412Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-07-20T21:45:44.027Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "product": "@ai-sdk/harness-codex",
          "vendor": "vercel",
          "versions": [
            {
              "status": "affected",
              "version": "\u003c 1.0.29"
            }
          ]
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "The `@ai-sdk/harness-opencode` tool is an HarnessV1 adapter backed by @openai/codex-sdk, which drives the codex command line interface. Prior to version 1.0.29, the tool relay authorizes requests from any process whose command line contains an allowed helper script path (the Codex CLI shim). This allows untrusted code executing in the sandbox to invoke arbitrary host-exposed tools, including secret lookups, deployment operations, and cloud API calls  without a corresponding model-authorized tool-call event. Exploitation requires a Linux environment (the vulnerable fallback checks `process.platform === \u0027linux\u0027` and reads `/proc`); an active harness session with one or more host-provided tools; and untrusted code executing in the sandbox (e.g. a malicious dependency, build script, or lifecycle hook) The fix in version 1.0.29 removes the process-path authorization fallback entirely. Relay requests are now only accepted after exact, short-lived, one-time authorization matching the tool name and input from a bridge-observed model event. Some workarounds are available. Do not run the Codex harness on untrusted repositories or with untrusted dependencies, and/or limit host-exposed tools to non-sensitive operations when working with untrusted code."
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "attackComplexity": "LOW",
            "attackRequirements": "NONE",
            "attackVector": "LOCAL",
            "baseScore": 6.3,
            "baseSeverity": "MEDIUM",
            "privilegesRequired": "NONE",
            "subAvailabilityImpact": "NONE",
            "subConfidentialityImpact": "HIGH",
            "subIntegrityImpact": "HIGH",
            "userInteraction": "NONE",
            "vectorString": "CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:N/VC:N/VI:L/VA:N/SC:H/SI:H/SA:N",
            "version": "4.0",
            "vulnAvailabilityImpact": "NONE",
            "vulnConfidentialityImpact": "NONE",
            "vulnIntegrityImpact": "LOW"
          }
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-863",
              "description": "CWE-863: Incorrect Authorization",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-07-20T20:27:41.349Z",
        "orgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
        "shortName": "GitHub_M"
      },
      "references": [
        {
          "name": "https://github.com/vercel/ai/security/advisories/GHSA-qw9h-448j-6rph",
          "tags": [
            "x_refsource_CONFIRM"
          ],
          "url": "https://github.com/vercel/ai/security/advisories/GHSA-qw9h-448j-6rph"
        },
        {
          "name": "https://github.com/vercel/ai/pull/17105",
          "tags": [
            "x_refsource_MISC"
          ],
          "url": "https://github.com/vercel/ai/pull/17105"
        }
      ],
      "source": {
        "advisory": "GHSA-qw9h-448j-6rph",
        "discovery": "UNKNOWN"
      },
      "title": "AI SDK Codex Harness Tool Relay Authorization Bypass"
    }
  },
  "cveMetadata": {
    "assignerOrgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
    "assignerShortName": "GitHub_M",
    "cveId": "CVE-2026-64650",
    "datePublished": "2026-07-20T20:27:41.349Z",
    "dateReserved": "2026-07-20T17:11:30.895Z",
    "dateUpdated": "2026-07-20T21:45:44.027Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2",
  "vulnerability-lookup:meta": {
    "epss": {
      "cve": "CVE-2026-64650",
      "date": "2026-10-27",
      "epss": "0.00174",
      "percentile": "0.06804"
    },
    "nvd": {
      "cve": {
        "affected": [
          {
            "affectedData": [
              {
                "product": "@ai-sdk/harness-codex",
                "vendor": "vercel",
                "versions": [
                  {
                    "status": "affected",
                    "version": "\u003c 1.0.29"
                  }
                ]
              }
            ],
            "source": "security-advisories@github.com"
          }
        ],
        "cveTags": [],
        "descriptions": [
          {
            "lang": "en",
            "value": "The `@ai-sdk/harness-opencode` tool is an HarnessV1 adapter backed by @openai/codex-sdk, which drives the codex command line interface. Prior to version 1.0.29, the tool relay authorizes requests from any process whose command line contains an allowed helper script path (the Codex CLI shim). This allows untrusted code executing in the sandbox to invoke arbitrary host-exposed tools, including secret lookups, deployment operations, and cloud API calls  without a corresponding model-authorized tool-call event. Exploitation requires a Linux environment (the vulnerable fallback checks `process.platform === \u0027linux\u0027` and reads `/proc`); an active harness session with one or more host-provided tools; and untrusted code executing in the sandbox (e.g. a malicious dependency, build script, or lifecycle hook) The fix in version 1.0.29 removes the process-path authorization fallback entirely. Relay requests are now only accepted after exact, short-lived, one-time authorization matching the tool name and input from a bridge-observed model event. Some workarounds are available. Do not run the Codex harness on untrusted repositories or with untrusted dependencies, and/or limit host-exposed tools to non-sensitive operations when working with untrusted code."
          },
          {
            "lang": "es",
            "value": "La herramienta \u201c@ai-sdk/harness-opencode\u201d es un adaptador HarnessV1 respaldado por @openai/codex-sdk, que impulsa la interfaz de l\u00ednea de comandos de codex. Antes de la versi\u00f3n 1.0.29, el rel\u00e9 de la herramienta autoriza solicitudes de cualquier proceso cuya l\u00ednea de comandos contenga una ruta de script auxiliar permitida (el shim de la CLI de Codex). Esto permite que c\u00f3digo no confiable ejecut\u00e1ndose en el sandbox invoque herramientas arbitrarias expuestas por el host, incluyendo b\u00fasquedas de secretos, operaciones de despliegue y llamadas a la API en la nube sin un evento de llamada a herramienta autorizado por el modelo correspondiente. La explotaci\u00f3n requiere un entorno Linux (la comprobaci\u00f3n de reserva vulnerable verifica \u201cprocess.platform === \u2019linux\u2019\u201d y lee \u201c/proc\u201d); una sesi\u00f3n de harness activa con una o m\u00e1s herramientas proporcionadas por el host; y c\u00f3digo no confiable ejecut\u00e1ndose en el sandbox (por ejemplo, una dependencia maliciosa, un script de compilaci\u00f3n o un hook de ciclo de vida). La correcci\u00f3n en la versi\u00f3n 1.0.29 elimina por completo la reserva de autorizaci\u00f3n por ruta de proceso. Las solicitudes de rel\u00e9 ahora solo se aceptan despu\u00e9s de una autorizaci\u00f3n exacta, de corta duraci\u00f3n y de un solo uso que coincida con el nombre de la herramienta y la entrada de un evento de modelo observado por el puente. Algunas soluciones alternativas est\u00e1n disponibles. No ejecute el harness de Codex en repositorios no confiables o con dependencias no confiables, y/o limite las herramientas expuestas por el host a operaciones no sensibles cuando trabaje con c\u00f3digo no confiable."
          }
        ],
        "id": "CVE-2026-64650",
        "lastModified": "2026-10-24T20:10:00.137",
        "metrics": {
          "cvssMetricV40": [
            {
              "cvssData": {
                "Automatable": "NOT_DEFINED",
                "Recovery": "NOT_DEFINED",
                "Safety": "NOT_DEFINED",
                "attackComplexity": "LOW",
                "attackRequirements": "NONE",
                "attackVector": "LOCAL",
                "availabilityRequirement": "NOT_DEFINED",
                "baseScore": 6.3,
                "baseSeverity": "MEDIUM",
                "confidentialityRequirement": "NOT_DEFINED",
                "exploitMaturity": "NOT_DEFINED",
                "integrityRequirement": "NOT_DEFINED",
                "modifiedAttackComplexity": "NOT_DEFINED",
                "modifiedAttackRequirements": "NOT_DEFINED",
                "modifiedAttackVector": "NOT_DEFINED",
                "modifiedPrivilegesRequired": "NOT_DEFINED",
                "modifiedSubAvailabilityImpact": "NOT_DEFINED",
                "modifiedSubConfidentialityImpact": "NOT_DEFINED",
                "modifiedSubIntegrityImpact": "NOT_DEFINED",
                "modifiedUserInteraction": "NOT_DEFINED",
                "modifiedVulnAvailabilityImpact": "NOT_DEFINED",
                "modifiedVulnConfidentialityImpact": "NOT_DEFINED",
                "modifiedVulnIntegrityImpact": "NOT_DEFINED",
                "privilegesRequired": "NONE",
                "providerUrgency": "NOT_DEFINED",
                "subAvailabilityImpact": "NONE",
                "subConfidentialityImpact": "HIGH",
                "subIntegrityImpact": "HIGH",
                "userInteraction": "NONE",
                "valueDensity": "NOT_DEFINED",
                "vectorString": "CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:N/VC:N/VI:L/VA:N/SC:H/SI:H/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X",
                "version": "4.0",
                "vulnAvailabilityImpact": "NONE",
                "vulnConfidentialityImpact": "NONE",
                "vulnIntegrityImpact": "LOW",
                "vulnerabilityResponseEffort": "NOT_DEFINED"
              },
              "source": "security-advisories@github.com",
              "type": "Secondary"
            }
          ],
          "ssvcV203": [
            {
              "source": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
              "ssvcData": {
                "id": "CVE-2026-64650",
                "options": [
                  {
                    "exploitation": "none"
                  },
                  {
                    "automatable": "no"
                  },
                  {
                    "technicalImpact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-07-20T21:45:38.151412Z",
                "version": "2.0.3"
              }
            }
          ]
        },
        "published": "2026-07-20T21:16:50.520",
        "references": [
          {
            "source": "security-advisories@github.com",
            "url": "https://github.com/vercel/ai/pull/17105"
          },
          {
            "source": "security-advisories@github.com",
            "url": "https://github.com/vercel/ai/security/advisories/GHSA-qw9h-448j-6rph"
          }
        ],
        "sourceIdentifier": "security-advisories@github.com",
        "vulnStatus": "Deferred",
        "weaknesses": [
          {
            "description": [
              {
                "lang": "en",
                "value": "CWE-863"
              }
            ],
            "source": "security-advisories@github.com",
            "type": "Secondary"
          }
        ]
      }
    },
    "vulnrichment": {
      "containers": {
        "adp": [
          {
            "metrics": [
              {
                "other": {
                  "content": {
                    "id": "CVE-2026-64650",
                    "options": [
                      {
                        "Exploitation": "none"
                      },
                      {
                        "Automatable": "no"
                      },
                      {
                        "Technical Impact": "partial"
                      }
                    ],
                    "role": "CISA Coordinator",
                    "timestamp": "2026-07-20T21:45:38.151412Z",
                    "version": "2.0.3"
                  },
                  "type": "ssvc"
                }
              }
            ],
            "providerMetadata": {
              "dateUpdated": "2026-07-20T21:45:40.564Z",
              "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
              "shortName": "CISA-ADP"
            },
            "title": "CISA ADP Vulnrichment"
          }
        ],
        "cna": {
          "affected": [
            {
              "product": "@ai-sdk/harness-codex",
              "vendor": "vercel",
              "versions": [
                {
                  "status": "affected",
                  "version": "\u003c 1.0.29"
                }
              ]
            }
          ],
          "descriptions": [
            {
              "lang": "en",
              "value": "The `@ai-sdk/harness-opencode` tool is an HarnessV1 adapter backed by @openai/codex-sdk, which drives the codex command line interface. Prior to version 1.0.29, the tool relay authorizes requests from any process whose command line contains an allowed helper script path (the Codex CLI shim). This allows untrusted code executing in the sandbox to invoke arbitrary host-exposed tools, including secret lookups, deployment operations, and cloud API calls  without a corresponding model-authorized tool-call event. Exploitation requires a Linux environment (the vulnerable fallback checks `process.platform === \u0027linux\u0027` and reads `/proc`); an active harness session with one or more host-provided tools; and untrusted code executing in the sandbox (e.g. a malicious dependency, build script, or lifecycle hook) The fix in version 1.0.29 removes the process-path authorization fallback entirely. Relay requests are now only accepted after exact, short-lived, one-time authorization matching the tool name and input from a bridge-observed model event. Some workarounds are available. Do not run the Codex harness on untrusted repositories or with untrusted dependencies, and/or limit host-exposed tools to non-sensitive operations when working with untrusted code."
            }
          ],
          "metrics": [
            {
              "cvssV4_0": {
                "attackComplexity": "LOW",
                "attackRequirements": "NONE",
                "attackVector": "LOCAL",
                "baseScore": 6.3,
                "baseSeverity": "MEDIUM",
                "privilegesRequired": "NONE",
                "subAvailabilityImpact": "NONE",
                "subConfidentialityImpact": "HIGH",
                "subIntegrityImpact": "HIGH",
                "userInteraction": "NONE",
                "vectorString": "CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:N/VC:N/VI:L/VA:N/SC:H/SI:H/SA:N",
                "version": "4.0",
                "vulnAvailabilityImpact": "NONE",
                "vulnConfidentialityImpact": "NONE",
                "vulnIntegrityImpact": "LOW"
              }
            }
          ],
          "problemTypes": [
            {
              "descriptions": [
                {
                  "cweId": "CWE-863",
                  "description": "CWE-863: Incorrect Authorization",
                  "lang": "en",
                  "type": "CWE"
                }
              ]
            }
          ],
          "providerMetadata": {
            "dateUpdated": "2026-07-20T20:27:41.349Z",
            "orgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
            "shortName": "GitHub_M"
          },
          "references": [
            {
              "name": "https://github.com/vercel/ai/security/advisories/GHSA-qw9h-448j-6rph",
              "tags": [
                "x_refsource_CONFIRM"
              ],
              "url": "https://github.com/vercel/ai/security/advisories/GHSA-qw9h-448j-6rph"
            },
            {
              "name": "https://github.com/vercel/ai/pull/17105",
              "tags": [
                "x_refsource_MISC"
              ],
              "url": "https://github.com/vercel/ai/pull/17105"
            }
          ],
          "source": {
            "advisory": "GHSA-qw9h-448j-6rph",
            "discovery": "UNKNOWN"
          },
          "title": "AI SDK Codex Harness Tool Relay Authorization Bypass"
        }
      },
      "cveMetadata": {
        "assignerOrgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
        "assignerShortName": "GitHub_M",
        "cveId": "CVE-2026-64650",
        "datePublished": "2026-07-20T20:27:41.349Z",
        "dateReserved": "2026-07-20T17:11:30.895Z",
        "dateUpdated": "2026-07-20T21:45:44.027Z",
        "state": "PUBLISHED"
      },
      "dataType": "CVE_RECORD",
      "dataVersion": "5.2"
    }
  }
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.




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.

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.
Vulnerability timeline

Sightings and recorded dates, grouped by day · UTC

JSON

Loading timeline…




    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.