{"id":"CVE-2026-56259","aliases":["EUVD-2026-43226","GHSA-f989-c77f-r2cq","GHSA-mw9f-q68p-364x","PYSEC-2026-3448"],"status":"published","title":"Crawl4AI - LLM Credential Exfiltration via base_url and Environment Variable Resolution","description":"Crawl4AI before 0.8.8 contains credential exfiltration vulnerabilities in the Docker API server that allow attackers to redirect LLM API calls to attacker-controlled endpoints and read arbitrary environment variables. Attackers can exploit the unauthenticated /md, /llm, and /llm/job endpoints by supplying a malicious base_url parameter and setting api_token to env:VARIABLE_NAME to exfiltrate provider API keys and server secrets including JWT SECRET_KEY for authentication bypass.","severity":{"version":"4.0","score":8.8,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:L/VA:N/SC:N/SI:N/SA:N","level":"high","assessor":"cna","from":"cve"},"epss":{"day":"2026-11-10","score":0.00433,"percentile":0.37933,"model":"v2026.06.15"},"facts":{"in_kev":false,"known_exploited":false,"public_exploit":false,"fix_available":true},"cwes":["CWE-200","CWE-522","CWE-918"],"published":"2026-06-16T21:00:31Z","modified":"2026-10-28T02:30:23Z","fixes":[{"ecosystem":"PyPI","package":"crawl4ai","fixed_in":["0.8.8"],"sources":["ghsa","osv"]}],"affected":[{"source":"ghsa","ecosystem":"PyPI","package":"crawl4ai","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"last_affected":"0.8.7"},{"fixed":"0.8.8"}]}]},{"source":"osv","ecosystem":"PyPI","package":"crawl4ai","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"0.8.8"}]}],"versions":["0.3.0","0.3.1","0.3.2","0.3.3","0.3.4","0.3.5","0.3.6","0.3.7","0.3.71","0.3.72","0.3.73","0.3.731","0.3.74","0.3.741","0.3.742","0.3.743","0.3.744","0.3.745","0.3.746","0.3.8","0.4.0","0.4.1","0.4.21","0.4.22","0.4.23","0.4.24","0.4.241","0.4.242","0.4.243","0.4.244","0.4.245","0.4.246","0.4.247","0.4.248","0.4.248b3","0.4.3b1","0.4.3b2","0.4.3b3","0.5.0","0.5.0.post1","0.5.0.post2","0.5.0.post3","0.5.0.post4","0.5.0.post5","0.5.0.post6","0.5.0.post7","0.5.0.post8","0.6.0","0.6.0rc1","0.6.1","0.6.2","0.6.3","0.7.0","0.7.1","0.7.2","0.7.3","0.7.4","0.7.5","0.7.6","0.7.7","0.7.8","0.8.0","0.8.5","0.8.6","0.8.7"]},{"source":"osv","ecosystem":"PyPI","package":"crawl4ai","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"0.8.8"}]}],"versions":["0.3.0","0.3.1","0.3.2","0.3.3","0.3.4","0.3.5","0.3.6","0.3.7","0.3.71","0.3.72","0.3.73","0.3.731","0.3.74","0.3.741","0.3.742","0.3.743","0.3.744","0.3.745","0.3.746","0.3.8","0.4.0","0.4.1","0.4.21","0.4.22","0.4.23","0.4.24","0.4.241","0.4.242","0.4.243","0.4.244","0.4.245","0.4.246","0.4.247","0.4.248","0.4.248b3","0.4.3b1","0.4.3b2","0.4.3b3","0.5.0","0.5.0.post1","0.5.0.post2","0.5.0.post3","0.5.0.post4","0.5.0.post5","0.5.0.post6","0.5.0.post7","0.5.0.post8","0.6.0","0.6.0rc1","0.6.1","0.6.2","0.6.3","0.7.0","0.7.1","0.7.2","0.7.3","0.7.4","0.7.5","0.7.6","0.7.7","0.7.8","0.8.0","0.8.5","0.8.6","0.8.7"]}],"products":[{"vendor":"crawl4ai","product":"crawl4ai"},{"vendor":"kidocode","product":"crawl4ai"},{"vendor":"unclecode","product":"crawl4ai"}],"references":[{"url":"https://github.com/unclecode/crawl4ai/security/advisories/GHSA-f989-c77f-r2cq","tags":["vendor-advisory","Vendor Advisory","Mitigation","advisory","web"]},{"url":"https://www.vulncheck.com/advisories/crawl4ai-llm-credential-exfiltration-via-base-url-and-environment-variable-resolution","tags":["third-party-advisory","Third Party Advisory","advisory"]},{"url":"https://github.com/advisories/GHSA-f989-c77f-r2cq"},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-56259","tags":["advisory"]},{"url":"https://github.com/advisories/GHSA-mw9f-q68p-364x"},{"url":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/56xxx/CVE-2026-56259.json","tags":["advisory"]},{"url":"https://github.com/unclecode/crawl4ai","tags":["package"]}],"sources":{"cve":[{"id":"CVE-2026-56259","title":"Crawl4AI - LLM Credential Exfiltration via base_url and Environment Variable Resolution","description":"Crawl4AI before 0.8.8 contains credential exfiltration vulnerabilities in the Docker API server that allow attackers to redirect LLM API calls to attacker-controlled endpoints and read arbitrary environment variables. Attackers can exploit the unauthenticated /md, /llm, and /llm/job endpoints by supplying a malicious base_url parameter and setting api_token to env:VARIABLE_NAME to exfiltrate provider API keys and server secrets including JWT SECRET_KEY for authentication bypass.","status":"published","published":"2026-07-12T12:06:30.919Z","modified":"2026-09-24T14:17:50.845Z","cvss":[{"version":"4.0","score":8.8,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:L/VA:N/SC:N/SI:N/SA:N","level":"high","assessor":"cna"},{"version":"3.1","score":8.2,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:L/A:N","level":"high","assessor":"cna"}],"cwes":["CWE-200"],"updated_at":"2026-11-02T13:37:02.780526Z"}],"euvd":[{"id":"EUVD-2026-43226","description":"Crawl4AI before 0.8.8 contains credential exfiltration vulnerabilities in the Docker API server that allow attackers to redirect LLM API calls to attacker-controlled endpoints and read arbitrary environment variables. Attackers can exploit the unauthenticated /md, /llm, and /llm/job endpoints by supplying a malicious base_url parameter and setting api_token to env:VARIABLE_NAME to exfiltrate provider API keys and server secrets including JWT SECRET_KEY for authentication bypass.","status":"published","published":"2026-07-12T12:06:30Z","modified":"2026-09-24T14:17:50Z","cvss":[{"version":"4.0","score":8.8,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:L/VA:N/SC:N/SI:N/SA:N","level":"high"}],"updated_at":"2026-11-02T17:27:58.186745Z"}],"ghsa":[{"id":"GHSA-f989-c77f-r2cq","title":"Crawl4AI: LLM credential exfiltration in Docker server via request base_url and env: token resolution","description":"### Summary\n\nThe Docker API server let a request control where LLM calls were sent and which environment variable an LLM token resolved from. Both could be abused to exfiltrate server-held secrets. The Docker API is unauthenticated by default.\n\n### Vector 1 - attacker base_url\n\n`/md`, `/llm`, and `/llm/job` accepted a `base_url` in the request and used it as the LLM endpoint while still attaching the server's configured provider API key. An attacker set `base_url` to a server they control and received the provider key (and any provider keys the server holds) in the inbound request.\n\n### Vector 2 - arbitrary environment variable read via `env:`\n\n`LLMConfig(api_token=\"env:NAME\")` resolved `NAME` from the server environment with `os.getenv`. Because request bodies were deserialized into `LLMConfig` (via a crawler config / extraction strategy), an attacker could set `api_token=\"env:SECRET_KEY\"` (or `env:REDIS_PASSWORD`, etc.) and, paired with an attacker `base_url`, exfiltrate that secret. Reading the server's `SECRET_KEY` enables forging authentication tokens.\n\n### Impact\n\nDisclosure of LLM provider API keys and other server secrets to an attacker-controlled endpoint; reading the JWT `SECRET_KEY` can lead to authentication bypass.\n\n### Fix\n\n- The LLM endpoints ignore a request-supplied `base_url`; the endpoint is always derived server-side from the provider name. The field is still accepted but no longer honored (no breaking 4xx).\n- `LLMConfig` refuses `env:` resolution of protected environment-variable names (names containing SECRET/PASSWORD/PRIVATE, prefixes CRAWL4AI*/AWS_SECRET*, and SECRET_KEY/REDIS_PASSWORD/TOKEN). Normal provider keys (e.g. OPENAI_API_KEY) are unaffected.\n\n### Workarounds\n\n- Upgrade to the patched version.\n- Enable authentication (`CRAWL4AI_API_TOKEN`).\n- Do not place sensitive secrets in the server environment alongside provider keys.\n\n### Credits\n\n- Geo ([geo-chen](https://github.com/geo-chen)) - reported the LLM credential exfiltration via request base_url.\n- Internal security audit (Crawl4AI maintainers) - the env: arbitrary-variable read.","status":"published","published":"2026-06-16T21:00:31Z","modified":"2026-06-16T21:00:31Z","cvss":[{"version":"3.1","score":8.2,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:L/A:N","level":"high"}],"cwes":["CWE-200","CWE-522","CWE-918"],"updated_at":"2026-11-02T14:40:28.438722Z"},{"id":"GHSA-mw9f-q68p-364x","title":"Crawl4AI before 0.8.8 contains credential exfiltration vulnerabilities in the Docker API server...","description":"Crawl4AI before 0.8.8 contains credential exfiltration vulnerabilities in the Docker API server that allow attackers to redirect LLM API calls to attacker-controlled endpoints and read arbitrary environment variables. Attackers can exploit the unauthenticated /md, /llm, and /llm/job endpoints by supplying a malicious base_url parameter and setting api_token to env:VARIABLE_NAME to exfiltrate provider API keys and server secrets including JWT SECRET_KEY for authentication bypass.","status":"published","published":"2026-07-12T12:31:47Z","modified":"2026-07-12T12:31:59Z","cvss":[{"version":"4.0","score":8.8,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:L/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"},{"version":"3.1","score":8.2,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:L/A:N","level":"high"}],"cwes":["CWE-200"],"updated_at":"2026-11-02T15:09:24.680615Z"}],"nvd":[{"id":"CVE-2026-56259","description":"Crawl4AI before 0.8.8 contains credential exfiltration vulnerabilities in the Docker API server that allow attackers to redirect LLM API calls to attacker-controlled endpoints and read arbitrary environment variables. Attackers can exploit the unauthenticated /md, /llm, and /llm/job endpoints by supplying a malicious base_url parameter and setting api_token to env:VARIABLE_NAME to exfiltrate provider API keys and server secrets including JWT SECRET_KEY for authentication bypass.","status":"published","published":"2026-07-12T12:16:45.013Z","modified":"2026-07-14T18:29:21.83Z","cvss":[{"version":"4.0","score":8.8,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:L/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":8.2,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:L/A:N","level":"high","assessor":"secondary"}],"cwes":["CWE-200"],"updated_at":"2026-11-02T14:12:28.771746Z"}],"osv":[{"id":"CVE-2026-56259","title":"Crawl4AI - LLM Credential Exfiltration via base_url and Environment Variable Resolution","description":"Crawl4AI before 0.8.8 contains credential exfiltration vulnerabilities in the Docker API server that allow attackers to redirect LLM API calls to attacker-controlled endpoints and read arbitrary environment variables. Attackers can exploit the unauthenticated /md, /llm, and /llm/job endpoints by supplying a malicious base_url parameter and setting api_token to env:VARIABLE_NAME to exfiltrate provider API keys and server secrets including JWT SECRET_KEY for authentication bypass.","status":"published","published":"2026-07-12T12:06:30.919Z","modified":"2026-10-28T02:30:23.350221056Z","cvss":[{"version":"4.0","score":8.8,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:L/VA:N/SC:N/SI:N/SA:N","level":"high"}],"cwes":["CWE-200"],"updated_at":"2026-11-02T15:45:11.609219Z"},{"id":"GHSA-f989-c77f-r2cq","title":"Crawl4AI: LLM credential exfiltration in Docker server via request base_url and env: token resolution","description":"### Summary\n\nThe Docker API server let a request control where LLM calls were sent and which environment variable an LLM token resolved from. Both could be abused to exfiltrate server-held secrets. The Docker API is unauthenticated by default.\n\n### Vector 1 - attacker base_url\n\n`/md`, `/llm`, and `/llm/job` accepted a `base_url` in the request and used it as the LLM endpoint while still attaching the server's configured provider API key. An attacker set `base_url` to a server they control and received the provider key (and any provider keys the server holds) in the inbound request.\n\n### Vector 2 - arbitrary environment variable read via `env:`\n\n`LLMConfig(api_token=\"env:NAME\")` resolved `NAME` from the server environment with `os.getenv`. Because request bodies were deserialized into `LLMConfig` (via a crawler config / extraction strategy), an attacker could set `api_token=\"env:SECRET_KEY\"` (or `env:REDIS_PASSWORD`, etc.) and, paired with an attacker `base_url`, exfiltrate that secret. Reading the server's `SECRET_KEY` enables forging authentication tokens.\n\n### Impact\n\nDisclosure of LLM provider API keys and other server secrets to an attacker-controlled endpoint; reading the JWT `SECRET_KEY` can lead to authentication bypass.\n\n### Fix\n\n- The LLM endpoints ignore a request-supplied `base_url`; the endpoint is always derived server-side from the provider name. The field is still accepted but no longer honored (no breaking 4xx).\n- `LLMConfig` refuses `env:` resolution of protected environment-variable names (names containing SECRET/PASSWORD/PRIVATE, prefixes CRAWL4AI*/AWS_SECRET*, and SECRET_KEY/REDIS_PASSWORD/TOKEN). Normal provider keys (e.g. OPENAI_API_KEY) are unaffected.\n\n### Workarounds\n\n- Upgrade to the patched version.\n- Enable authentication (`CRAWL4AI_API_TOKEN`).\n- Do not place sensitive secrets in the server environment alongside provider keys.\n\n### Credits\n\n- Geo ([geo-chen](https://github.com/geo-chen)) - reported the LLM credential exfiltration via request base_url.\n- Internal security audit (Crawl4AI maintainers) - the env: arbitrary-variable read.","status":"published","published":"2026-06-16T21:00:31Z","modified":"2026-07-15T20:11:30.37162885Z","cvss":[{"version":"3.1","score":8.2,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:L/A:N","level":"high"}],"cwes":["CWE-200","CWE-522","CWE-918"],"updated_at":"2026-11-02T15:46:46.006631Z"},{"id":"PYSEC-2026-3448","description":"Crawl4AI before 0.8.8 contains credential exfiltration vulnerabilities in the Docker API server that allow attackers to redirect LLM API calls to attacker-controlled endpoints and read arbitrary environment variables. Attackers can exploit the unauthenticated /md, /llm, and /llm/job endpoints by supplying a malicious base_url parameter and setting api_token to env:VARIABLE_NAME to exfiltrate provider API keys and server secrets including JWT SECRET_KEY for authentication bypass.","status":"published","published":"2026-07-12T12:16:45.013Z","modified":"2026-07-15T20:00:06.248966718Z","cvss":[{"version":"4.0","score":8.8,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:L/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"}],"updated_at":"2026-11-02T15:52:56.847083Z"}]}}