llamaCVEVulnerability Intelligence
/
Advisory · GitHub · PyPI · openssl-encrypt

GHSA-QJR2-X6MR-8XGF

Imported identity email printed unsanitized lets ANSI escapes forge the fingerprint verification line

Parche disponible
Descargar PNG JSON
CVSS
—
medium
Fuente: —
ExplotaciónSin evidenciaNo figura en CISA KEV
Exploits públicos0Ninguno conocido
Parche1 versiónPrimer parche a 1 d de la divulgación
Exposición99 dDesde la divulgación, 12 ago 2026, 7:19 a. m.
Detección98 dDesde la divulgación hasta su ingreso al índice
Timeline

Ciclo de vida de la vulnerabilidad

Hoy
Divulgación
Publicada
Explotación
Remediación
Parche 1.4.9
Monitoreo
Detectada +1
12 ago 26
13 ago 26
18 nov 26
19 nov 26
12 ago 26
6 sept 26
1 oct 26
26 oct 26
19 nov 26
FechaFaseEventoDetalle
Divulgación Publicada GHSA-QJR2-X6MR-8XGF divulgada
Remediación Parche 1.4.9 openssl-encrypt · 1 d desde divulgación
Monitoreo Detectada Ingresó al índice
Monitoreo Verificada Última verificación contra las fuentes
Resumen
Descripción del advisory
Texto original · GitHub · en inglés
Summary

`Identity.import_public` validated the identity `name` but accepted the `email` field of an imported identity document completely raw, and the CLI printed it to the terminal unsanitized (`identity import` success output, `identity list`, `identity show`) — directly above the `Fingerprint:` line. A JSON string may contain `\u001b` escape sequences (the JSON security validator rejects only *literal* control characters, whose escaped source text is printable), so a crafted bundle such as

{ "email": "x\u001b[1A\u001b[2KFingerprint: <a fingerprint the victim trusts>\u001b[1A" }

can move the cursor and overwrite the genuine `Fingerprint:` line with an attacker-chosen one. Out-of-band fingerprint comparison is the only authenticity mechanism the identity design has, so forging that line attacks the trust model directly.

Security review of the fix found the same class live on every other identity display surface, including one reachable by a fully remote attacker: the keyserver TOFU trust prompt prints a fetched bundle's `name`/`email`/`created_at` raw around the very fingerprint it tells the user to verify out of band (`created_at` is printed *after* the fingerprint and was completely unconstrained). The bundle's self-signature is no defence — it verifies against the signing key shipped *in* the bundle. Further instances: `identity create --email` (producer side — the value is exported verbatim and uploaded, attacking other users' prompts), stored identity files read back from disk (an `--identity-store` directory feeds `list`/`show` and the TOFU key-change warning with no import-time check), a signature sidecar leaking through three fields (`signer_fingerprint` printed on the unknown-signer error path before any cryptographic check; `algorithm` echoed by the unsupported-algorithm error pre-verification; and the display-only `component` name — deliberately excluded from the signed payload, so any tamperer can rewrite it on a *valid* signature — printed inside the GOOD-signature verdict block), a keyserver's raw HTTP error body (unbounded remote text) reaching the terminal through exception messages, and error paths interpolating a rejected name or `--alias` verbatim.

Impact

An attacker who can deliver an identity bundle to the victim — the normal contact-exchange flow (`--file`, the GUI's paste field) or a keyserver response — can forge the fingerprint readout used for out-of-band verification, enabling exactly the key substitution the TOFU pinning ceremony exists to surface. Reachable in all released versions.

Fix
  • Identity metadata is validated at every boundary: `import_public`, `Identity.generate`, `Identity.load`, and `PublicKeyBundle` validate `email` (string, ≤ 320 chars, no control characters), `fingerprint` (colon-separated lowercase hex — the only shape the tool has ever written), and `created_at` (≤ 64 chars, no controls); `parse_signature` format-validates `signer_fingerprint`, `algorithm`, and every `component` name; keyserver HTTP error bodies are truncated and sanitized; `identity.json` is read bounded, explicitly UTF-8, and through the JSON security validator.
  • A display sanitizer escapes C0 controls, DEL, the C1 range (one-byte CSI introducers included), backslash (output unambiguity), and the bidi/format controls (honoured by VTE/Kitty, able to visually reverse an email within its line) at every terminal display of these fields: the identity CLI, the keyserver trust prompt and `keyserver search`, the TOFU key-change warning, the verified-signature sender line, and every identity CLI error path. Escaping rather than stripping keeps the evidence visible.
  • Contacts and cached keyserver bundles imported by earlier versions with crafted fields are neutralized at display time by the same sanitizer.
Versiones

Rangos afectados

Afectada Corregida Última
openssl-encrypt≤ 1.4.8
1.4.8
1.4.9
Remediación

Estado del parche por paquete

Paquete / productoEstadoCorregida enDesde divulgaciónÚltima versión
pypiopenssl-encrypt
Corregida 1.4.9 1 d 1.4.9 no afectada
Versiones corregidas1.4.9
Explotación

Arsenal público

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

Referencias
Advisoryhttps://github.com/jahlives/openssl_encrypt/security/advisories/GHSA-qjr2-x6mr-8xgf

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