Questa pagina è stata tradotta automaticamente. Leggi l’originale in inglese.

Component

core/security/nemo-guardrails

NVIDIA NeMo Guardrails' server as a guard in a model's path: a service speaking the OpenAI chat-completions API that checks each request against its input rail, forwards what it admits to the upstream model API in infra.upstream_endpoint, and returns the upstream's answer, streamed or not. A request the rail refuses is answered by the guard itself with an ordinary completion, and nothing of it reaches the upstream. The rail sends the latest user message to the detector wired on the detector socket and refuses the request when the detector reports one of values.blocked_labels at or above values.threshold; a detector that cannot be reached, or answers with an error, refuses the request too. It judges the latest user message alone. Its verdict is a probabilistic classifier's: it misses some attacks, refuses some legitimate requests and is known to be evadable. Only a flow in its configuration, redefining the product's own, keeps a request body from switching the rail off, and that closure has to be proved again for any image or flow it was not proved on. NVIDIA's NemoGuard models add rails on the content_safety and jailbreak_detection sockets; the jailbreak rail admits a request when its classifier is unreachable. It authenticates no caller, and with no credential attached to its upstream_key slot a caller can reach the upstream on a credential of its own. The Release supplies the image in infra.image and the catalog names none: TrustyAI's published build carries NLTK's punkt tokenizer model, whose licence is unresolved upstream, and is published only under a floating tag, whose superseded digests stop resolving. Nothing renders until the Release sets `infra.image`.

Provenienza

UfficialeFirmato da un’identità di firma che questo catalogo riconosce come editore ufficiale.
Digest del contenuto
sha256:f871e845d92afd93e2d6d8c7fc6a9b33f86abbcfe1730cc89aab9d44fb3f2721
Pubblicato
2026-11-10T10:13:39Z
Identità di firma
https://token.actions.githubusercontent.com/https://github.com/aggrega-ai/aggrega-hub/.github/workflows/catalog-publish-sign.yml@refs/heads/main

Firma

  • Ufficialeapplication/vnd.dev.sigstore.bundle.v0.3+json

    https://token.actions.githubusercontent.com/https://github.com/aggrega-ai/aggrega-hub/.github/workflows/catalog-publish-sign.yml@refs/heads/main

    sha256:4a2b09d9b13c5c2efc29928cc9bd5abc2304b72043e199a4b9ffddab6eb8b06e

Riferimento a questa versione

oci://hub.aggrega.ai/component/core/security/nemo-guardrails:1.0.0

Parametri dello sviluppatore

Valori che lo sviluppatore imposta nell’Aggregate.

NomeTipoPredefinitoObbligatorioDescrizione
thresholdnumber (min 0, max 1)0.5noProbability at or above which a label the detector reports refuses the request. Measured with the detector's default model, which scores most texts close to 0 or close to 1: at 0.5 the guard refuses plain attempts to override a model's instructions; at 0.001 it also refuses an override respelled in leetspeak, which the model scores near 0.008, and with it some ordinary requests the model scores low but not near zero. At either value it refuses the ordinary requests the model scores close to 1, and serves attacks it scores close to 0. Those two numbers hold for that model only, and are measured again for any other.
blocked_labelsarray (min 1)INJECTIONnoLabels of the detector's classifier that refuse a request when reported at or above the threshold. The default is the label the detector's default model gives an attempt to override a model's instructions; another model labels by its own vocabulary, which this list follows. It may not be empty, because a guard refusing on no label forwards everything.

Dimensionamento dell’operatore

Dimensionamento dell’infrastruttura che l’operatore regola per ogni ambiente.

NomeTipoPredefinitoObbligatorioDescrizione
portinteger (min 1024, max 65535)8000noPort the guard listens on and its Service accepts callers on. Most operators should omit this and let the Recipe assign the default (8000), the port the published image listens on unless told otherwise.
upstream_endpointstring—sìBase URL of the OpenAI-compatible model API the guard forwards the requests it admits to, with its API path, for example http://model-server:8000/v1. A deterministic pointer string the guard resolves when it forwards; the compiler never contacts it. The request is forwarded under the model name the caller asked for.
imagestring—sìImage the guard runs, supplied by the Release; the catalog names none. The Recipes run the published build's own start-up and paths, so the image is TrustyAI's NeMo Guardrails server or a copy of it. That build carries a data package whose licence is unresolved upstream and is tagged only as latest, under which the registry stops serving a build once the tag moves on, so a Release that must keep one build mirrors it and names the copy here. The guard's refusal of a request-body switch was proved on one build, and is proved again for any other image.
computeobject—no—

Contratto di output

Valori che questo componente pubblica verso i consumatori attraverso la topologia uses.

NomeTipoPredefinitoObbligatorioDescrizione
hoststring—sìIn-cluster DNS name of the Service fronting the guard. Every caller reaching this name can send requests through the guard, so it is an address for the component routing model traffic through it rather than one to publish.
portinteger—sìTCP port that Service accepts callers on. It follows the port an operator set in the Release, so a consumer reads it rather than assuming a default.
endpointstring—sìBase URL of the OpenAI-compatible API the guard serves, with the host, the port and the API path already composed. A consumer hands it to a client library as its base URL, and sends chat completions to it.
protocolstring—noURL scheme to reach `host` and `port` over, for a consumer composing its own address rather than reading `endpoint`.
thresholdnumber—sìvalues.threshold, the probability at or above which a label the detector reports refuses the request.

Socket uses

I socket di dipendenza attraverso cui questo componente si connette. La destinazione di ogni connessione collegata attraverso un socket deve soddisfare il suo selettore di etichette, e ogni connessione può portare un aggancio negli slot che il socket dichiara. Uno slot obbligatorio deve essere riempito da ogni connessione che l’Aggregate o la Release collega attraverso il socket.

detector

Selettore
security.aggrega.ai/text-detection=detector
Iniezione
—

Questo socket non dichiara slot di aggancio, quindi una connessione attraverso di esso non porta alcun aggancio.

content_safety

Selettore
component.aggrega.ai/category=ai-model, model.aggrega.ai/api=openai
Iniezione
—

Questo socket non dichiara slot di aggancio, quindi una connessione attraverso di esso non porta alcun aggancio.

jailbreak_detection

Selettore
component.aggrega.ai/category=security
Iniezione
—

Questo socket non dichiara slot di aggancio, quindi una connessione attraverso di esso non porta alcun aggancio.

Slot di aggancio

Punti di associazione con nome in cui una Feature si aggancia a questo componente. Il selettore di uno slot limita quale Feature può riempirlo e la sua lista inject indica come può essere consegnato il materiale della Feature. Uno slot obbligatorio deve essere associato su ogni istanza, da un aggancio sull’istanza o a un ambito superiore, e uno slot sensibile è dichiarato per gestire materiale segreto. Una Feature può anche agganciarsi con un nome che nessuno slot dichiara.

SlotSelettoreObbligatorioIniezioneSensibileDescrizione
upstream_keyfeature.aggrega.ai/kind=secretnoenvsìOptional credential the guard presents to the upstream model API as a bearer token. Attached, every forwarded request carries it, whatever the caller sends. Unattached, the guard holds no upstream credential, which suits an upstream that asks for none: the product then sends a fixed placeholder in its place, which an upstream requiring a key refuses, or passes a caller's X-Authorization header on as the credential when the caller sends one.

Elenco consentito di auto-iniezione

Nessuna pipe di auto-iniezione dichiarata.

Relazioni

Archi tipizzati da e verso altre entità pubblicate. Un riferimento senza una destinazione pubblicata viene mostrato come arco interrotto anziché come link.

Usa

Usato da

Riferimento al concetto Component →

Il catalogo Aggrega — un indice aperto di componenti, funzionalità, ricette e dei plugin che li renderizzano.

Realizzato con il design system condiviso di Aggrega.

Privacy