{"api_version":"1","generated_at":"2026-07-21T17:50:18+00:00","cve":"CVE-2026-64651","urls":{"html":"https://cve.report/CVE-2026-64651","api":"https://cve.report/api/cve/CVE-2026-64651.json","docs":"https://cve.report/api","cve_org":"https://www.cve.org/CVERecord?id=CVE-2026-64651","nvd":"https://nvd.nist.gov/vuln/detail/CVE-2026-64651"},"summary":{"title":"AI SDK OpenCode Harness Tool Relay Authorization Bypass","description":"The `@ai-sdk/harness-opencode` tool connects HarnessAgent to OpenCode through a sandboxed bridge. Prior to version 1.0.28, the tool relay authorizes requests from any process whose command line contains an allowed helper script path (`host-tool-mcp.mjs`). 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 1.0.28 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 OpenCode harness on untrusted repositories or with untrusted dependencies, and/or limit host-exposed tools to non-sensitive operations when working with untrusted code.","state":"PUBLISHED","assigner":"GitHub_M","published_at":"2026-07-20 21:16:50","updated_at":"2026-07-21 13:17:18"},"problem_types":["CWE-863","CWE-863 CWE-863: Incorrect Authorization"],"metrics":[{"version":"4.0","source":"security-advisories@github.com","type":"Secondary","score":"6.3","severity":"MEDIUM","vector":"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","data":{"version":"4.0","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","baseScore":6.3,"baseSeverity":"MEDIUM","attackVector":"LOCAL","attackComplexity":"LOW","attackRequirements":"NONE","privilegesRequired":"NONE","userInteraction":"NONE","vulnConfidentialityImpact":"NONE","vulnIntegrityImpact":"LOW","vulnAvailabilityImpact":"NONE","subConfidentialityImpact":"HIGH","subIntegrityImpact":"HIGH","subAvailabilityImpact":"NONE","exploitMaturity":"NOT_DEFINED","confidentialityRequirement":"NOT_DEFINED","integrityRequirement":"NOT_DEFINED","availabilityRequirement":"NOT_DEFINED","modifiedAttackVector":"NOT_DEFINED","modifiedAttackComplexity":"NOT_DEFINED","modifiedAttackRequirements":"NOT_DEFINED","modifiedPrivilegesRequired":"NOT_DEFINED","modifiedUserInteraction":"NOT_DEFINED","modifiedVulnConfidentialityImpact":"NOT_DEFINED","modifiedVulnIntegrityImpact":"NOT_DEFINED","modifiedVulnAvailabilityImpact":"NOT_DEFINED","modifiedSubConfidentialityImpact":"NOT_DEFINED","modifiedSubIntegrityImpact":"NOT_DEFINED","modifiedSubAvailabilityImpact":"NOT_DEFINED","Safety":"NOT_DEFINED","Automatable":"NOT_DEFINED","Recovery":"NOT_DEFINED","valueDensity":"NOT_DEFINED","vulnerabilityResponseEffort":"NOT_DEFINED","providerUrgency":"NOT_DEFINED"}},{"version":"4.0","source":"CNA","type":"DECLARED","score":"6.3","severity":"MEDIUM","vector":"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","data":{"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"}}],"references":[{"url":"https://github.com/vercel/ai/security/advisories/GHSA-g48p-5rr5-8rgq","name":"https://github.com/vercel/ai/security/advisories/GHSA-g48p-5rr5-8rgq","refsource":"security-advisories@github.com","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://github.com/vercel/ai/pull/17105","name":"https://github.com/vercel/ai/pull/17105","refsource":"security-advisories@github.com","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://www.cve.org/CVERecord?id=CVE-2026-64651","name":"CVE Program record","refsource":"CVE.ORG","tags":["canonical"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-64651","name":"NVD vulnerability detail","refsource":"NVD","tags":["canonical","analysis"]}],"affected":[{"source":"CNA","vendor":"vercel","product":"@ai-sdk/harness-opencode","version":"affected < 1.0.29","platforms":[]}],"timeline":[],"solutions":[],"workarounds":[],"exploits":[],"credits":[],"nvd_cpes":[],"vendor_comments":[],"enrichments":{"kev":null,"epss":null,"legacy_qids":[]},"source_records":{"cve_program":{"containers":{"adp":[{"metrics":[{"other":{"content":{"id":"CVE-2026-64651","options":[{"Exploitation":"none"},{"Automatable":"no"},{"Technical Impact":"partial"}],"role":"CISA Coordinator","timestamp":"2026-07-21T12:49:24.395782Z","version":"2.0.3"},"type":"ssvc"}}],"providerMetadata":{"dateUpdated":"2026-07-21T12:49:47.977Z","orgId":"134c704f-9b21-4f2e-91b3-4a467353bcc0","shortName":"CISA-ADP"},"title":"CISA ADP Vulnrichment"}],"cna":{"affected":[{"product":"@ai-sdk/harness-opencode","vendor":"vercel","versions":[{"status":"affected","version":"< 1.0.29"}]}],"descriptions":[{"lang":"en","value":"The `@ai-sdk/harness-opencode` tool connects HarnessAgent to OpenCode through a sandboxed bridge. Prior to version 1.0.28, the tool relay authorizes requests from any process whose command line contains an allowed helper script path (`host-tool-mcp.mjs`). 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 1.0.28 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 OpenCode 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":"2026-07-20T20:26:07.497Z","orgId":"a0819718-46f1-4df5-94e2-005712e83aaa","shortName":"GitHub_M"},"references":[{"name":"https://github.com/vercel/ai/security/advisories/GHSA-g48p-5rr5-8rgq","tags":["x_refsource_CONFIRM"],"url":"https://github.com/vercel/ai/security/advisories/GHSA-g48p-5rr5-8rgq"},{"name":"https://github.com/vercel/ai/pull/17105","tags":["x_refsource_MISC"],"url":"https://github.com/vercel/ai/pull/17105"}],"source":{"advisory":"GHSA-g48p-5rr5-8rgq","discovery":"UNKNOWN"},"title":"AI SDK OpenCode Harness Tool Relay Authorization Bypass"}},"cveMetadata":{"assignerOrgId":"a0819718-46f1-4df5-94e2-005712e83aaa","assignerShortName":"GitHub_M","cveId":"CVE-2026-64651","datePublished":"2026-07-20T20:26:07.497Z","dateReserved":"2026-07-20T17:11:30.896Z","dateUpdated":"2026-07-21T12:49:47.977Z","state":"PUBLISHED"},"dataType":"CVE_RECORD","dataVersion":"5.2"},"nvd":{"publishedDate":"2026-07-20 21:16:50","lastModifiedDate":"2026-07-21 13:17:18","problem_types":["CWE-863","CWE-863 CWE-863: Incorrect Authorization"],"metrics":{"cvssMetricV40":[{"source":"security-advisories@github.com","type":"Secondary","cvssData":{"version":"4.0","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","baseScore":6.3,"baseSeverity":"MEDIUM","attackVector":"LOCAL","attackComplexity":"LOW","attackRequirements":"NONE","privilegesRequired":"NONE","userInteraction":"NONE","vulnConfidentialityImpact":"NONE","vulnIntegrityImpact":"LOW","vulnAvailabilityImpact":"NONE","subConfidentialityImpact":"HIGH","subIntegrityImpact":"HIGH","subAvailabilityImpact":"NONE","exploitMaturity":"NOT_DEFINED","confidentialityRequirement":"NOT_DEFINED","integrityRequirement":"NOT_DEFINED","availabilityRequirement":"NOT_DEFINED","modifiedAttackVector":"NOT_DEFINED","modifiedAttackComplexity":"NOT_DEFINED","modifiedAttackRequirements":"NOT_DEFINED","modifiedPrivilegesRequired":"NOT_DEFINED","modifiedUserInteraction":"NOT_DEFINED","modifiedVulnConfidentialityImpact":"NOT_DEFINED","modifiedVulnIntegrityImpact":"NOT_DEFINED","modifiedVulnAvailabilityImpact":"NOT_DEFINED","modifiedSubConfidentialityImpact":"NOT_DEFINED","modifiedSubIntegrityImpact":"NOT_DEFINED","modifiedSubAvailabilityImpact":"NOT_DEFINED","Safety":"NOT_DEFINED","Automatable":"NOT_DEFINED","Recovery":"NOT_DEFINED","valueDensity":"NOT_DEFINED","vulnerabilityResponseEffort":"NOT_DEFINED","providerUrgency":"NOT_DEFINED"}}],"ssvcV203":[{"source":"134c704f-9b21-4f2e-91b3-4a467353bcc0","ssvcData":{"timestamp":"2026-07-21T12:49:24.395782Z","id":"CVE-2026-64651","options":[{"exploitation":"none"},{"automatable":"no"},{"technicalImpact":"partial"}],"role":"CISA Coordinator","version":"2.0.3"}}]},"configurations":[]},"legacy_mitre":{"record":{"CveYear":"2026","CveId":"64651","Ordinal":"1","Title":"AI SDK OpenCode Harness Tool Relay Authorization Bypass","CVE":"CVE-2026-64651","Year":"2026"},"notes":[{"CveYear":"2026","CveId":"64651","Ordinal":"1","NoteData":"The `@ai-sdk/harness-opencode` tool connects HarnessAgent to OpenCode through a sandboxed bridge. Prior to version 1.0.28, the tool relay authorizes requests from any process whose command line contains an allowed helper script path (`host-tool-mcp.mjs`). 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 1.0.28 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 OpenCode harness on untrusted repositories or with untrusted dependencies, and/or limit host-exposed tools to non-sensitive operations when working with untrusted code.","Type":"Description","Title":"AI SDK OpenCode Harness Tool Relay Authorization Bypass"}]}}}