llamaCVEVulnerability Intelligence
/
Advisory · GitHub · Go · code.vikunja.io/veans

GHSA-3RRQ-H284-5VPX

Repo config redirects VEANS_TOKEN to an attacker-chosen server

Sin parche publicado
Descargar PNG JSON
CVSS 3.1
5.9
Media
CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:L/A:N Fuente: GitHub
ExplotaciónSin evidenciaNo figura en CISA KEV
Exploits públicos0Ninguno conocido
ParcheSin parcheEl fabricante no ha publicado corrección
Exposición28 dDesde la divulgación, 22 oct 2026, 5:00 a. m.
Detección27 dDesde la divulgación hasta su ingreso al índice
Timeline

Ciclo de vida de la vulnerabilidad

Hoy
Divulgación
Publicada
Explotación
Remediación
Monitoreo
Detectada +1
22 oct 26
18 nov 26
20 nov 26
22 oct 26
29 oct 26
5 nov 26
12 nov 26
20 nov 26
FechaFaseEventoDetalle
Divulgación Publicada GHSA-3RRQ-H284-5VPX divulgada · CVSS 5.9 (GitHub)
Monitoreo Detectada Ingresó al índice
Monitoreo Verificada Última verificación contra las fuentes
Resumen
Descripción del advisory
Texto original · GitHub · en inglés

Affects all veans versions (veans has no tagged release yet).

## Summary
veans takes the Vikunja server URL from the nearest `.veans.yml`, walking up from the working directory. When `VEANS_TOKEN` is set, veans sends it as a Bearer token to whatever server that file names. Running any veans command in an untrusted repo, or CI running veans on PR code, hands the token to the repo author's server.

## Details
- `EnvBackend.Get` returns the env token regardless of server/account (`veans/internal/credentials/env.go:29`).
- `config.Find` walks up from cwd (`veans/internal/config/config.go:102`), and `loadRuntime` passes `cfg.Server` to the client (`veans/internal/commands/runtime.go:57,63`).
- The client adds the Bearer header with no origin check (`veans/internal/client/client.go:106,128`).
- `prime` triggers such a request (`veans/internal/commands/prime.go:78`), and the agent hooks run it on every session start.
- Keyring and file tokens are keyed by server (`veans/internal/credentials/store.go:57`), so only the env token leaks.

## Impact
The attacker gets the victim's Vikunja API token and whatever access it carries on the real server. Requires `VEANS_TOKEN` in the environment (typically CI) and veans running in an attacker-controlled checkout.

## Fix
Add a trust gate: one origin check in the client/runtime setup (`veans/internal/commands/runtime.go:57-63`, `veans/internal/commands/login.go:69`), run before any env token is attached, any human credential is sent (password, token, or opening the OAuth browser), or any remote metadata is fetched for the prompt.

  • Trust records live in global XDG state, never repo-local (an attacker repo could ship a pre-approved one). They are keyed by normalized origin (scheme + host + port) plus repo identity, so an `https` record doesn't cover a downgraded `http` origin.
  • Interactive commands show the normalized origin and ask once; "yes" records it. `veans init` with a user-typed server records trust without asking.
  • Non-interactive runs (the SessionStart/PreCompact hook running `prime`, CI) never prompt and fail closed.
  • In CI, a `VEANS_SERVER` env var whose origin matches `cfg.Server` counts as a trust record.

Attach `VEANS_TOKEN` only to a trusted origin: an interactive trust record, or, in CI (`veans/internal/credentials/env.go:21-23`), a matching `VEANS_SERVER`. With neither, veans refuses and doesn't prompt.

_Found via internal security review (csf #2, f2 #3, 03 #12, 04 #9, deep #13)._

Folded in: `veans login` sends human credentials to a repo-selected server (was GHSA-2mwp-p5v5-3gcm)

Same root cause, same trust gate. A teammate runs `veans login` in a cloned repo, as veans itself tells them to (`veans/internal/commands/runtime.go:60`). Their password, TOTP code or token goes to the server named in the repo's `.veans.yml`; in OAuth mode they land on a sign-in page hosted by that server. `login.go:56,69` (`config.Find` → `client.New(cfg.Server, "")`) has no origin confirmation, the password prompt is generic (`auth.go:143`) before the POST (`:152`), and the server is named only after success (`login.go:109`). The gate must run before `AcquireHumanToken`, which covers password, token and OAuth modes.

Gate requirements found in verification: normalize with `net/url` and reject userinfo (`client.New` string-concatenates `BaseURL`, so `https://trusted.example@evil.example/` currently goes to evil.example); run it before `store.Get` in `loadRuntime`, because `backfillFrontendURL` (`runtime.go:72-74`) sends an authenticated request inside `loadRuntime` itself; store the record under a fixed `~/.config/veans/` path with no XDG env override.

Vector CVSS 3.1

Desglose de métricas

Opción resaltada = valor del vector · izquierda = peor caso · GitHub
Vector de ataqueAV:N
RedAdyacenteLocalFísica
ComplejidadAC:H
BajaAlta
PrivilegiosPR:N
NingunoBajosAltos
InteracciónUI:R
NingunaRequerida
AlcanceS:U
CambiadoSin cambio
ConfidencialidadC:H
AltoBajoNinguno
IntegridadI:L
AltoBajoNinguno
DisponibilidadA:N
AltoBajoNinguno
Versiones

Rangos afectados

Afectada Corregida
code.vikunja.io/veans≤ 0.0.0
0.0.0
Remediación

Estado del parche por paquete

Paquete / productoEstadoCorregida enDesde divulgaciónÚltima versión
Gocode.vikunja.io/veans
Sin parche — — —
Explotación

Arsenal público

Sin exploits, módulos ni PoC públicos conocidos a la fecha de verificación.

Referencias
Advisoryhttps://github.com/go-vikunja/vikunja/security/advisories/GHSA-3rrq-h284-5vpx

Las referencias externas se muestran para copiar: cve.llama.ec no lo redirige a otros sitios.