GCVE Workshop - 22 September 2026 (14:00-18:00), Luxembourg Before The Vulnopticon Conference - Registration

GHSA-5X9F-6VG5-QG4M

Vulnerability from github – Published: 2026-06-05 15:25 – Updated: 2026-06-05 15:25
VLAI
Summary
Omni has a TOCTOU race condition that allows multiple concurrent uses of a single-use SAML session token
Details

Summary

SAML.getSession (internal/pkg/auth/interceptor/saml.go) checks the Used flag on a SAMLAssertion resource and then marks it used in two separate state operations. Because the check and the update are not atomic, concurrent requests carrying the same saml-session token can both observe Used == false, both pass validation, and both return a successful authentication context. An attacker who obtains a valid saml-session token can exploit this window to authenticate as the token's owner multiple times, defeating the one-time-use guarantee.

Severity

  • Attack Vector: Local: the attacker needs to either be able to intercept the local, unencrypted traffic or needs access to user's browser.
  • Attack Complexity: High: the attacker must first obtain a valid saml-session token belonging to the victim (requires a separate interception step; the token is ephemeral and single-use by design).
  • Privileges Required: None: no Omni account is required to carry out the race once the session token is in hand.
  • User Interaction: Required: the victim must initiate a SAML authentication flow to produce the session token that the attacker intercepts.
  • Scope: Unchanged: the impact stays within Omni's authorization boundary.
  • Confidentiality Impact: High: successful exploitation authenticates the attacker as the victim's email identity, granting read access to any resource accessible to that identity.
  • Integrity Impact: High: the attacker can confirm one or more public keys under the victim's identity (via ConfirmPublicKey), establishing persistent access credentials tied to the victim's account.
  • Availability Impact: High: if the attacker can successfully perform the attack and if the victim is a privileged Omni user, e.g., an Omni Operator or Admin, they can take Omni down.

Impact

  • Session replay: A stolen saml-session token can be used more than once, defeating its single-use guarantee.
  • Multiple public key confirmations: An attacker who steals the session can confirm N attacker-controlled public keys under the victim's identity in a single stolen session window, creating N persistent long-lived API credentials tied to the victim's account.
  • Authentication as victim: Any gRPC endpoint gated by the SAML interceptor can be reached as the victim's email identity during the race window.
  • Audit log pollution: Each raced call generates an audit entry attributed to the victim's email, obscuring the attacker's actions.

Credit

This vulnerability was discovered and reported by bugbunny.ai.

Severity
7.0 (High)
CVSS:3.1/AV:L/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H
Show details on source website
To clipboard 

{
  "affected": [
    {
      "package": {
        "ecosystem": "Go",
        "name": "github.com/siderolabs/omni"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "1.6.6"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    },
    {
      "package": {
        "ecosystem": "Go",
        "name": "github.com/siderolabs/omni"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "1.7.0"
            },
            {
              "fixed": "1.7.3"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-45720"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-294",
      "CWE-367"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-06-05T15:25:28Z",
    "nvd_published_at": null,
    "severity": "HIGH"
  },
  "details": "## Summary\n\n`SAML.getSession` (`internal/pkg/auth/interceptor/saml.go`) checks the `Used` flag on a `SAMLAssertion` resource and then marks it used in two separate state operations. Because the check and the update are not atomic, concurrent requests carrying the same `saml-session` token can both observe `Used == false`, both pass validation, and both return a successful authentication context. An attacker who obtains a valid `saml-session` token can exploit this window to authenticate as the token\u0027s owner multiple times, defeating the one-time-use guarantee.\n\n## Severity\n\n- **Attack Vector:** Local: the attacker needs to either be able to intercept the local, unencrypted traffic or needs access to user\u0027s browser.\n- **Attack Complexity:** High: the attacker must first obtain a valid `saml-session` token belonging to the victim (requires a separate interception step; the token is ephemeral and single-use by design).\n- **Privileges Required:** None: no Omni account is required to carry out the race once the session token is in hand.\n- **User Interaction:** Required: the victim must initiate a SAML authentication flow to produce the session token that the attacker intercepts.\n- **Scope:** Unchanged: the impact stays within Omni\u0027s authorization boundary.\n- **Confidentiality Impact:** High: successful exploitation authenticates the attacker as the victim\u0027s email identity, granting read access to any resource accessible to that identity.\n- **Integrity Impact:** High: the attacker can confirm one or more public keys under the victim\u0027s identity (via `ConfirmPublicKey`), establishing persistent access credentials tied to the victim\u0027s account.\n- **Availability Impact:** High: if the attacker can successfully perform the attack and if the victim is a privileged Omni user, e.g., an Omni Operator or Admin, they can take Omni down.\n\n## Impact\n\n- **Session replay**: A stolen `saml-session` token can be used more than once, defeating its single-use guarantee.\n- **Multiple public key confirmations**: An attacker who steals the session can confirm N attacker-controlled public keys under the victim\u0027s identity in a single stolen session window, creating N persistent long-lived API credentials tied to the victim\u0027s account.\n- **Authentication as victim**: Any gRPC endpoint gated by the SAML interceptor can be reached as the victim\u0027s email identity during the race window.\n- **Audit log pollution**: Each raced call generates an audit entry attributed to the victim\u0027s email, obscuring the attacker\u0027s actions.\n\n## Credit\n\nThis vulnerability was discovered and reported by [bugbunny.ai](https://bugbunny.ai).",
  "id": "GHSA-5x9f-6vg5-qg4m",
  "modified": "2026-06-05T15:25:28Z",
  "published": "2026-06-05T15:25:28Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/siderolabs/omni/security/advisories/GHSA-5x9f-6vg5-qg4m"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/siderolabs/omni"
    },
    {
      "type": "WEB",
      "url": "https://github.com/siderolabs/omni/releases/tag/v1.6.6"
    },
    {
      "type": "WEB",
      "url": "https://github.com/siderolabs/omni/releases/tag/v1.7.3"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:L/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H",
      "type": "CVSS_V3"
    }
  ],
  "summary": "Omni has a TOCTOU race condition that allows multiple concurrent uses of a single-use SAML session token"
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.





Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

Sightings

Author Source Type Date Other

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.


Detection rules are retrieved from Rulezet.


Related by attack behaviour

Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.