GHSA-33RQ-M5X2-FVGF

Vulnerability from github – Published: 2026-02-17 21:37 – Updated: 2026-03-11 20:41
VLAI
Summary
OpenClaw Twitch allowFrom is not enforced in optional plugin, unauthorized chat users can trigger agent pipeline
Details

Summary

In the optional Twitch channel plugin (extensions/twitch), allowFrom is documented as a hard allowlist of Twitch user IDs, but it was not enforced as a hard gate. If allowedRoles is unset or empty, the access control path defaulted to allow, so any Twitch user who could mention the bot could reach the agent dispatch pipeline.

Scope note: This only affects deployments that installed and enabled the Twitch plugin. Core OpenClaw installs that do not install/enable the Twitch plugin are not impacted.

Affected Packages / Versions

  • Package: openclaw (npm)
  • Affected: >= 2026.1.29, < 2026.2.1
  • Fixed: >= 2026.2.1

Details

Affected component: Twitch plugin access control (extensions/twitch/src/access-control.ts).

Problematic logic in checkTwitchAccessControl():

  • When allowFrom was configured, the code returned allowed: true for members but did not return allowed: false for non-members, so execution fell through.
  • If allowedRoles was unset or empty, the function returned allowed: true by default, even when allowFrom was configured.

Proof of Concept (PoC)

  1. Install and enable the Twitch plugin.
  2. Configure an allowFrom list, but do not set allowedRoles (or set it to an empty list).
  3. From a different Twitch account whose user ID is NOT in allowFrom, send a message that mentions the bot (for example @<botname> hello).
  4. Observe the message is processed and can trigger agent dispatch/replies despite not being allowlisted.

Impact

Authorization bypass for operators who relied on allowFrom to restrict who can invoke the bot in Twitch chat. Depending on configuration (tools, routing, model costs), this could lead to unintended actions/responses and resource or cost exhaustion.

Fix Commit(s)

  • 8c7901c984866a776eb59662dc9d8b028de4f0d0

Workaround

Upgrade to openclaw >= 2026.2.1.

Thanks @MegaManSec (https://joshua.hu) of AISLE Research Team for reporting.

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

{
  "affected": [
    {
      "package": {
        "ecosystem": "npm",
        "name": "openclaw"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "2026.1.29"
            },
            {
              "fixed": "2026.2.1"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-28448"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-285"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-02-17T21:37:55Z",
    "nvd_published_at": "2026-03-05T22:16:16Z",
    "severity": "HIGH"
  },
  "details": "### Summary\n\nIn the optional Twitch channel plugin (`extensions/twitch`), `allowFrom` is documented as a hard allowlist of Twitch user IDs, but it was not enforced as a hard gate. If `allowedRoles` is unset or empty, the access control path defaulted to allow, so any Twitch user who could mention the bot could reach the agent dispatch pipeline.\n\n**Scope note:** This only affects deployments that installed and enabled the Twitch plugin. Core OpenClaw installs that do not install/enable the Twitch plugin are not impacted.\n\n### Affected Packages / Versions\n\n- Package: `openclaw` (npm)\n- Affected: `\u003e= 2026.1.29, \u003c 2026.2.1`\n- Fixed: `\u003e= 2026.2.1`\n\n### Details\n\nAffected component: Twitch plugin access control (`extensions/twitch/src/access-control.ts`).\n\nProblematic logic in `checkTwitchAccessControl()`:\n\n- When `allowFrom` was configured, the code returned `allowed: true` for members but did not return `allowed: false` for non-members, so execution fell through.\n- If `allowedRoles` was unset or empty, the function returned `allowed: true` by default, even when `allowFrom` was configured.\n\n### Proof of Concept (PoC)\n\n1. Install and enable the Twitch plugin.\n2. Configure an `allowFrom` list, but do not set `allowedRoles` (or set it to an empty list).\n3. From a different Twitch account whose user ID is NOT in `allowFrom`, send a message that mentions the bot (for example `@\u003cbotname\u003e hello`).\n4. Observe the message is processed and can trigger agent dispatch/replies despite not being allowlisted.\n\n### Impact\n\nAuthorization bypass for operators who relied on `allowFrom` to restrict who can invoke the bot in Twitch chat. Depending on configuration (tools, routing, model costs), this could lead to unintended actions/responses and resource or cost exhaustion.\n\n### Fix Commit(s)\n\n- `8c7901c984866a776eb59662dc9d8b028de4f0d0`\n\n### Workaround\n\nUpgrade to `openclaw \u003e= 2026.2.1`.\n\nThanks @MegaManSec (https://joshua.hu) of [AISLE Research Team](https://aisle.com/) for reporting.",
  "id": "GHSA-33rq-m5x2-fvgf",
  "modified": "2026-03-11T20:41:24Z",
  "published": "2026-02-17T21:37:55Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/openclaw/openclaw/security/advisories/GHSA-33rq-m5x2-fvgf"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-28448"
    },
    {
      "type": "WEB",
      "url": "https://github.com/openclaw/openclaw/commit/8c7901c984866a776eb59662dc9d8b028de4f0d0"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/openclaw/openclaw"
    },
    {
      "type": "WEB",
      "url": "https://github.com/openclaw/openclaw/releases/tag/v2026.2.1"
    },
    {
      "type": "WEB",
      "url": "https://www.vulncheck.com/advisories/openclaw-authorization-bypass-in-twitch-plugin-allowfrom-access-control"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L",
      "type": "CVSS_V3"
    }
  ],
  "summary": "OpenClaw Twitch allowFrom is not enforced in optional plugin, unauthorized chat users can trigger agent pipeline"
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.




Drag nodes to arrange them. Double-click a related vulnerability to open it, or a source to visit its link. The Physics button on the left tunes the layout and the simulation; the legend at the bottom left hides or shows nodes by kind and by sighting type.

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.
Vulnerability timeline

Sightings and recorded dates, grouped by day · UTC

JSON
How durations are calculated

Following the VTEM framework (Black Hat USA 2015), time to patch starts when the vendor or a coordinator was informed, total time to patch at discovery, and disclosure to patch at public disclosure; vendor response time runs from the report to its acknowledgement. Time to exploit and time unpatched start at the earliest discovery or report, falling back to public disclosure. Unpatched time ends at the first recorded patch, or the calculation time when no patch is recorded. Exploitation to patch shows the interval between those two events; a published proof of concept is a public exploit, not exploitation.

Each result shows its source dates. A negative interval means the end event came first; when date-only inputs leave the order open, the result says so. Date-only inputs are approximate. The earliest evidence wins: KEV report dates and sightings are labelled as such. Records describing another vulnerability are listed but never calculated from, and a PoC alone does not establish exploitation. Missing patch evidence does not prove that no patch exists. Filters below change the event list only.

Loading timeline…




    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.