Sanaan Fayaz Wani

teableio/teable

GHSA-w677-p6hx-85vw

Share password bypass: the share access token is never bound to the share it is presented for, so a token minted for one share opens any other password-protected share (plus cleartext password disclosure)

GHSA-w677-p6hx-85vw

7.1

High

CVSS 3.1 base

  1. 2026-09-05Reported privately
  2. 38 days later2026-10-13Published by the maintainer

What to do

If you run teable before release.2026-10-12T12-45-53Z.3331, upgrade to release.2026-10-12T12-45-53Z.3331 or later.

Anyone who can reach it over the network, with a low-privilege account and no user interaction, can read sensitive data and change some data.Read from the CVSS 3.1 vector

CVSS 3.1 metrics

To exploit it

Attack vector
Network
Attack complexity
Low
Privileges required
Low
User interaction
None

What it gets

Confidentiality
High
Integrity
Low
Availability
None
Scope
Unchanged

Open this vector in the FIRST calculator ↗

Weakness
CWE-287, CWE-522, CWE-863 Improper Authentication · Insufficiently Protected Credentials · Incorrect Authorization
CVE
not assigned
Vector
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:L/A:N
Also in teable
GHSA-jvm6-86c5-f8j8 6.5
2026-09-05Reported privately
38 days
2026-10-13Published by the maintainer

My analysis of this finding: root cause, the vulnerable code, the impact and the fix, in the site’s format. The official advisory of record is linked above.

Affected versions

Package Ecosystem Vulnerable Fixed in
teableio/teable other < release.2026-10-12T12-45-53Z.3331 release.2026-10-12T12-45-53Z.3331

Summary

Teable lets you protect a shared view or a shared base with an optional password. I found that the share-auth JWT is never bound to the share it is presented for. The password is verified against the shareId carried inside the token payload, while the data served comes from the shareId in the request URL. Those are allowed to be two different shares. Any authenticated user can mint a legitimate token for a password-protected share they own, relabel it under a victim share’s cookie name, and read the victim’s protected content without ever knowing the victim’s password. Both the view-share and the base-share paths carry the flaw. This is an authorization bypass through a user-controlled key (CWE-639).

Where it lives

The Passport strategy at apps/nestjs-backend/src/features/share/strategies/jwt.strategy.ts:

public static fromAuthCookieAsToken(req: Request): string | null {  const shareId = req.params.shareId || (req.headers['tea-share-id'] as string);  const cookieObj = cookie.parse(req.headers.cookie ?? '');  return cookieObj?.[shareId] ?? null;              // requested share picks the COOKIE NAME} async validate(req, payload: IJwtShareInfo) {  const { shareId, password } = payload;            // the PAYLOAD picks what gets authenticated  const authShareId = await this.shareAuthService.authShareView(shareId, password, req.useV2 === true);  if (!authShareId) throw new UnauthorizedException();  return authShareId;                               // no payload.shareId === req.params.shareId check}

The bugthe requested share only picks which cookie is readThe bugthe token payload picks which share gets authenticatedThe bugno check that the token's share matches the request

authShareView validates the password of the share named in the payload. By the time validate runs, the guard has already resolved req.shareInfo to the requested (victim) share, and every handler serves from req.shareInfo. So the credential checked and the data returned come from two different shares.

The base-share leg is identical in apps/nestjs-backend/src/features/base-share/strategies/jwt.strategy.ts, calling authBaseShare(shareId, password) with the same missing bind.

It is an oversight, not a design choice: the sibling header-token validators bind correctly. Both validateBaseSharePasswordToken (permission.service.ts:590) and validateShareViewPasswordToken (:1054) do exactly the check the two strategies omit:

if (payload.shareId !== shareId) { return false; }

Impact

I verified this end to end against ghcr.io/teableio/teable:latest with a base on the current default v2Enabled=true. A non-admin attacker minted a token for their own password share shrau2hPeaD7RoEL23A, then hit a victim share whose password VictimSecretPASS123 they never learned:

Control  GET /api/share/shrbkg5yCzDfPtNHjB1/view/records            -> 401 (no token)
Exploit  GET /api/share/shrbkg5yCzDfPtNHjB1/view/records
         Cookie: shrbkg5yCzDfPtNHjB1=<attacker token for shrau2...> -> 200
         {"records":[{"id":"rec89zy140BEKDH2F6h",
           "fields":{"fldvuv2MTJ6gqELp2Kq":"TOP-SECRET-SALARY-999999"},...}]}

The share password is the only control protecting a link whose URL has leaked or been forwarded. Any authenticated holder of a victim share link reads the full protected view, and against a base share can export and copy the entire password-protected base.

Fix

In both validate methods, reject when the token’s shareId does not match the requested shareId, mirroring the header validators:

const requestedShareId = req.params.shareId || req.headers['tea-share-id'];if (payload.shareId !== requestedShareId) throw new UnauthorizedException();

The fixrejects tokens minted for a different share

The base-share strategy needs passReqToCallback: true first so it can see the request at all.

The record

Weakness
CWE-287 Improper Authentication
CWE-522 Insufficiently Protected Credentials
CWE-863 Incorrect Authorization
Affected
teable before release.2026-10-12T12-45-53Z.3331
Fixed in
release.2026-10-12T12-45-53Z.3331
Disclosure
2026-09-05 Reported privately
38 days later2026-10-13 Published
Advisory
GHSA-w677-p6hx-85vw
Also in teable
GHSA-jvm6-86c5-f8j8 6.5

About this writeup

Sanaan Fayaz Wani (GitHub sfwani) reported this vulnerability to the teable maintainers under coordinated disclosure and is credited as a reporter in GHSA-w677-p6hx-85vw, published 2026-10-13.

All published findings: advisory index.