CVE-2025-68665: @langchain/core LangChain Serialization Injection Vulnerability Enables Vulnerability

{
  "schema_version" : "1.4.0",
  "id" : "GHSA-r399-636x-v7f6",
  "modified" : "2025-12-24T01:08:11Z",
  "published" : "2025-12-23T20:08:48Z",
  "aliases" : [ "CVE-2025-68665" ],
  "summary" : "LangChain serialization injection vulnerability enables secret extraction",
  "details" : "## Context\n\nA serialization injection vulnerability exists in LangChain JS's `toJSON()` method (and subsequently when string-ifying objects using `JSON.stringify()`. The method did not escape objects with `'lc'` keys when serializing free-form data in kwargs. The `'lc'` key is used internally by LangChain to mark serialized objects. When user-controlled data contains this key structure, it is treated as a legitimate LangChain object during deserialization rather than plain user data.\n\n### Attack surface\n\nThe core vulnerability was in `Serializable.toJSON()`: this method failed to escape user-controlled objects containing `'lc'` keys within kwargs (e.g., `additional_kwargs`, `metadata`, `response_metadata`). When this unescaped data was later deserialized via `load()`, the injected structures were treated as legitimate LangChain objects rather than plain user data.\n\nThis escaping bug enabled several attack vectors:\n\n1. **Injection via user data**: Malicious LangChain object structures could be injected through user-controlled fields like `metadata`, `additional_kwargs`, or `response_metadata`\n2. **Secret extraction**: Injected secret structures could extract environment variables when `secretsFromEnv` was enabled (which had no explicit default, effectively defaulting to `true` behavior)\n3. **Class instantiation via import maps**: Injected constructor structures could instantiate any class available in the provided import maps with attacker-controlled parameters\n\n**Note on import maps:** Classes must be explicitly included in import maps to be instantiatable. The core import map includes standard types (messages, prompts, documents), and users can extend this via `importMap` and `optionalImportsMap` options. This architecture naturally limits the attack surface—an `allowedObjects` parameter is not necessary because users control which classes are available through the import maps they provide.\n\n**Security hardening:** This patch fixes the escaping bug in `toJSON()` and introduces new restrictive defaults in `load()`: `secretsFromEnv` now explicitly defaults to `false`, and a `maxDepth` parameter protects against DoS via deeply nested structures. JSDoc security warnings have been added to all import map options.\n\n## Who is affected?\n\nApplications are vulnerable if they:\n\n1. **Serialize untrusted data via `JSON.stringify()` on Serializable objects, then deserialize with `load()`** — Trusting your own serialization output makes you vulnerable if user-controlled data (e.g., from LLM responses, metadata fields, or user inputs) contains `'lc'` key structures.\n2. **Deserialize untrusted data with `load()`** — Directly deserializing untrusted data that may contain injected `'lc'` structures.\n3. **Use LangGraph checkpoints** — Checkpoint serialization/deserialization paths may be affected.\n\nThe most common attack vector is through **LLM response fields** like `additional_kwargs` or `response_metadata`, which can be controlled via prompt injection and then serialized/deserialized in streaming operations.\n\n## Impact\n\nAttackers who control serialized data can extract environment variable secrets by injecting `{\"lc\": 1, \"type\": \"secret\", \"id\": [\"ENV_VAR\"]}` to load environment variables during deserialization (when `secretsFromEnv: true`). They can also instantiate classes with controlled parameters by injecting constructor structures to instantiate any class within the provided import maps with attacker-controlled parameters, potentially triggering side effects such as network calls or file operations.\n\nKey severity factors:\n\n- Affects the serialization path—applications trusting their own serialization output are vulnerable\n- Enables secret extraction when combined with `secretsFromEnv: true`\n- LLM responses in `additional_kwargs` can be controlled via prompt injection\n\n## Exploit example\n\n```typescript\nimport { load } from \"@langchain/core/load\";\n\n// Attacker injects secret structure into user-controlled data\nconst attackerPayload = JSON.stringify({\n  user_data: {\n    lc: 1,\n    type: \"secret\",\n    id: [\"OPENAI_API_KEY\"],\n  },\n});\n\nprocess.env.OPENAI_API_KEY = \"sk-secret-key-12345\";\n\n// With secretsFromEnv: true, the secret is extracted\nconst deserialized = await load(attackerPayload, { secretsFromEnv: true });\n\nconsole.log(deserialized.user_data); // \"sk-secret-key-12345\" - SECRET LEAKED!\n```\n\n## Security hardening changes\n\nThis patch introduces the following changes to `load()`:\n\n1. **`secretsFromEnv` default changed to `false`**: Disables automatic secret loading from environment variables. Secrets not found in `secretsMap` now throw an error instead of being loaded from `process.env`. This fail-safe behavior ensures missing secrets are caught immediately rather than silently continuing with `null`.\n2. **New `maxDepth` parameter** (defaults to `50`): Protects against denial-of-service attacks via deeply nested JSON structures that could cause stack overflow.\n3. **Escape mechanism in `toJSON()`**: User-controlled objects containing `'lc'` keys are now wrapped in `{\"__lc_escaped__\": {...}}` during serialization and unwrapped as plain data during deserialization.\n4. **JSDoc security warnings**: All import map options (`importMap`, `optionalImportsMap`, `optionalImportEntrypoints`) now include security warnings about never populating them from user input.\n\n## Migration guide\n\n### No changes needed for most users\n\nIf you're deserializing standard LangChain types (messages, documents, prompts) using the core import map, your code will work without changes:\n\n```typescript\nimport { load } from \"@langchain/core/load\";\n\n// Works with default settings\nconst obj = await load(serializedData);\n```\n\n### For secrets from environment\n\n`secretsFromEnv` now defaults to `false`, and missing secrets throw an error. If you need to load secrets:\n\n```typescript\nimport { load } from \"@langchain/core/load\";\n\n// Provide secrets explicitly (recommended)\nconst obj = await load(serializedData, {\n  secretsMap: { OPENAI_API_KEY: process.env.OPENAI_API_KEY },\n});\n\n// Or explicitly opt-in to load from env (only use with trusted data)\nconst obj = await load(serializedData, { secretsFromEnv: true });\n```\n\n> **Warning:** Only enable `secretsFromEnv` if you trust the serialized data. Untrusted data could extract any environment variable.\n\n> **Note:** If a secret reference is encountered but not found in `secretsMap` (and `secretsFromEnv` is `false` or the secret is not in the environment), an error is thrown. This fail-safe behavior ensures you're aware of missing secrets rather than silently receiving `null` values.\n\n### For deeply nested structures\n\nIf you have legitimate deeply nested data that exceeds the default depth limit of 50:\n\n```typescript\nimport { load } from \"@langchain/core/load\";\n\nconst obj = await load(serializedData, { maxDepth: 100 });\n```\n\n### For custom import maps\n\nIf you provide custom import maps, ensure they only contain trusted modules:\n\n```typescript\nimport { load } from \"@langchain/core/load\";\nimport * as myModule from \"./my-trusted-module\";\n\n// GOOD - explicitly include only trusted modules\nconst obj = await load(serializedData, {\n  importMap: { my_module: myModule },\n});\n\n// BAD - never populate from user input\nconst obj = await load(serializedData, {\n  importMap: userProvidedImports, // DANGEROUS!\n});\n```",
  "severity" : [ {
    "type" : "CVSS_V3",
    "score" : "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:N/A:N"
  } ],
  "affected" : [ {
    "package" : {
      "ecosystem" : "npm",
      "name" : "@langchain/core"
    },
    "ranges" : [ {
      "type" : "ECOSYSTEM",
      "events" : [ {
        "introduced" : "1.0.0"
      }, {
        "fixed" : "1.1.8"
      } ]
    } ]
  }, {
    "package" : {
      "ecosystem" : "npm",
      "name" : "@langchain/core"
    },
    "ranges" : [ {
      "type" : "ECOSYSTEM",
      "events" : [ {
        "introduced" : "0"
      }, {
        "fixed" : "0.3.80"
      } ]
    } ]
  }, {
    "package" : {
      "ecosystem" : "npm",
      "name" : "langchain"
    },
    "ranges" : [ {
      "type" : "ECOSYSTEM",
      "events" : [ {
        "introduced" : "1.0.0"
      }, {
        "fixed" : "1.2.3"
      } ]
    } ]
  }, {
    "package" : {
      "ecosystem" : "npm",
      "name" : "langchain"
    },
    "ranges" : [ {
      "type" : "ECOSYSTEM",
      "events" : [ {
        "introduced" : "0"
      }, {
        "fixed" : "0.3.37"
      } ]
    } ]
  } ],
  "references" : [ {
    "type" : "WEB",
    "url" : "https://github.com/langchain-ai/langchainjs/security/advisories/GHSA-r399-636x-v7f6"
  }, {
    "type" : "ADVISORY",
    "url" : "https://nvd.nist.gov/vuln/detail/CVE-2025-68665"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/langchain-ai/langchainjs/commit/e5063f9c6e9989ea067dfdff39262b9e7b6aba62"
  }, {
    "type" : "PACKAGE",
    "url" : "https://github.com/langchain-ai/langchainjs"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain%2Fcore%401.1.8"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/langchain-ai/langchainjs/releases/tag/langchain%401.2.3"
  } ],
  "database_specific" : {
    "cwe_ids" : [ "CWE-502" ],
    "severity" : "HIGH",
    "github_reviewed" : true,
    "github_reviewed_at" : "2025-12-23T20:08:48Z",
    "nvd_published_at" : "2025-12-23T23:15:45Z"
  }
}