Skip to content

Text Format string map parsing can mutate returned map object prototype

Moderate
dcodeIO published GHSA-jfj6-75fj-8934 Jul 4, 2026

Package

npm protobufjs (npm)

Affected versions

>=8.2.0, <=8.6.4

Patched versions

8.6.5

Description

Summary

The protobuf.js text format extension parsed string-keyed map entries using ordinary property assignment. A text-format map entry with key __proto__ could therefore change the prototype of the returned map object instead of creating an own map entry.

This affected the optional Text Format extension. Other protobufjs input paths, including binary decode, fromObject, and ProtoJSON conversion, are not affected.

Impact

An attacker who can provide protobuf text-format input parsed by an application using protobufjs/ext/textformat may be able to create message objects whose string-keyed map fields have attacker-controlled prototypes.

This is per-object prototype mutation, not global Object.prototype pollution. Impact depends on downstream application logic treating inherited properties as meaningful, for example by using in, truthiness checks, or direct property access on parsed map objects instead of own-property checks.

Applications that do not parse untrusted Text Format input, or that do not use inherited properties from parsed map objects in security-relevant logic, are not directly affected.

Preconditions

  • The application must parse attacker-controlled protobuf Text Format input with protobufjs/ext/textformat.
  • The target schema must contain a string-keyed map field.
  • The crafted input must provide a map entry with key __proto__.
  • Downstream application logic must treat inherited properties on the returned map object as meaningful for impact beyond malformed output.

Workarounds

Upgrade to protobufjs 8.6.5 or newer.

If immediate upgrade is not possible, do not parse untrusted protobuf Text Format input with affected versions. Applications can also reject string map keys named __proto__ before or during Text Format parsing, and should use own-property checks such as Object.hasOwnProperty.call(map, key) when consuming parsed map objects.

Severity

Moderate

CVSS overall score

This score calculates overall vulnerability severity from 0 to 10 and is based on the Common Vulnerability Scoring System (CVSS).
/ 10

CVSS v3 base metrics

Attack vector
Network
Attack complexity
High
Privileges required
None
User interaction
None
Scope
Unchanged
Confidentiality
Low
Integrity
Low
Availability
None

CVSS v3 base metrics

Attack vector: More severe the more the remote (logically and physically) an attacker can be in order to exploit the vulnerability.
Attack complexity: More severe for the least complex attacks.
Privileges required: More severe if no privileges are required.
User interaction: More severe when no user interaction is required.
Scope: More severe when a scope change occurs, e.g. one vulnerable component impacts resources in components beyond its security scope.
Confidentiality: More severe when loss of data confidentiality is highest, measuring the level of data access available to an unauthorized user.
Integrity: More severe when loss of data integrity is the highest, measuring the consequence of data modification possible by an unauthorized user.
Availability: More severe when the loss of impacted component availability is highest.
CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:N

CVE ID

CVE-2026-59876

Weaknesses

Improperly Controlled Modification of Object Prototype Attributes ('Prototype Pollution')

The product receives input from an upstream component that specifies attributes that are to be initialized or updated in an object, but it does not properly control modifications of attributes of the object prototype. Learn more on MITRE.

Credits