Component
core/security/trustyai-huggingface-detector
TrustyAI's Hugging Face detector: a service that runs one text classifier from the Hugging Face Hub and reports what it detects in the texts a caller sends. For each text posted to its content route with a threshold, it answers with every label the classifier scores at or above it, never the classifier's first label, read as the one for harmless text. It changes nothing and forwards nothing: what to do about a detection is the caller's decision, and core/security/nemo-guardrails is one such caller. The classifier is values.model at values.model_revision. The default, protectai/deberta-v3-base-prompt-injection-v2 (Apache-2.0), labels each text SAFE or INJECTION, and its own model card states that the project which published it is archived and no longer maintained. Another model labels and scores differently, so a caller's labels and thresholds follow it. The verdict misses some attacks and flags some ordinary requests. The weights are fetched where the service runs, never at build time. It authenticates no caller and every text reaches it as sent, so it is published on an in-cluster address only. Its own metrics, in the Prometheus format at /metrics on the same port, are published as `metrics_endpoint`. The Release supplies the image in infra.image and the catalog names none: TrustyAI's published build carries NVIDIA's proprietary CUDA runtime, whose licence agreement limits its use to systems with NVIDIA GPUs, and is published only under a floating tag, whose superseded digests stop resolving. Nothing renders until the Release sets `infra.image`.
Procedencia
- Digest del contenido
- sha256:8f985fb76602947f3d871a7625d6614d76375528aa12e151d26a51559254d630
- Publicado
- 2026-11-10T10:13:59Z
- Identidad de firma
- https://token.actions.githubusercontent.com/https://github.com/aggrega-ai/aggrega-hub/.github/workflows/catalog-publish-sign.yml@refs/heads/main
Firma
- Oficialapplication/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:c34f4d9c71225ccd5e22a7d802c196dda140050af22eb57aa716bedbe73525f8
Referencia en esta versión
oci://hub.aggrega.ai/component/core/security/trustyai-huggingface-detector:1.0.0Parámetros del desarrollador
Valores que el desarrollador establece en el Aggregate.
| Nombre | Tipo | Predeterminado | Obligatorio | Descripción |
|---|---|---|---|---|
| model | string | protectai/deberta-v3-base-prompt-injection-v2 | no | Hugging Face model id of the sequence-classification model the service runs. The default is a prompt-injection classifier labelling each text SAFE or INJECTION; another model labels by its own vocabulary, which a caller's refusal labels must then follow. |
| model_revision | string | 90c9989b1a342275dd0d1a95aad283c04e075671 | no | Revision of values.model to fetch — a commit id, a tag or a branch. The default is the commit of the default model whose scores a caller's thresholds are measured against; set it together with values.model, and main follows the repository's newest revision. |
Dimensionamiento del operador
Dimensionamiento de la infraestructura que el operador ajusta para cada entorno.
| Nombre | Tipo | Predeterminado | Obligatorio | Descripción |
|---|---|---|---|---|
| port | integer (min 1024, max 65535) | 8000 | no | Port the service 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. |
| image | string | — | sí | Image the service and its weights-pull step run, supplied by the Release; the catalog names none. The Recipes run the published build's own server command, log configuration and Hugging Face command-line tool, so the image is TrustyAI's Hugging Face detector or a copy of it. That build carries NVIDIA's proprietary CUDA runtime, whose licence agreement limits its use to systems with NVIDIA GPUs, 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. |
| compute | object | — | no | — |
| model_cache | object | — | no | The volume the pull step writes the weights to and the server reads them from. |
Contrato de salida
Valores que este componente publica para los consumidores a través de la topología uses.
| Nombre | Tipo | Predeterminado | Obligatorio | Descripción |
|---|---|---|---|---|
| host | string | — | sí | In-cluster DNS name of the Service fronting the detector. Every caller reaching this name can submit text, so it is an address for the component that calls the service rather than one to publish. |
| port | integer | — | sí | TCP 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. |
| endpoint | string | — | sí | Base URL of the detector's HTTP API, with the host and the port already composed and no path. A caller posts the texts to check to the content route under it. |
| model | string | — | sí | Hugging Face model id of the classifier the service loaded, so a caller can record which model gave a verdict. |
| protocol | string | — | no | URL scheme to reach `host` and `port` over, for a consumer composing its own address rather than reading `endpoint`. |
| metrics_endpoint | string | — | no | The full address Prometheus scrapes this component's metrics at — scheme, host, port and path, `http://<host>:<port>/metrics` on the port callers reach. Empty where the component serves none reachable. |
Sockets uses
Los sockets de dependencia a través de los cuales se conecta este componente. El destino de cada conexión cableada a través de un socket debe satisfacer su selector de etiquetas, y cada conexión puede llevar un acoplamiento en los slots que declara el socket. Cada conexión que el Aggregate o la Release cablea a través del socket debe llenar los slots obligatorios.
No se declararon sockets uses.
Slots de acoplamiento
Puntos de enlace con nombre en los que una Feature se acopla a este componente. El selector de un slot restringe qué Feature puede llenarlo y su lista inject indica cómo puede entregarse el material de la Feature. Un slot obligatorio debe estar enlazado en cada instancia, mediante un acoplamiento en la instancia o en un ámbito superior, y un slot sensible se declara para manejar material secreto. Una Feature también puede acoplarse con un nombre que ningún slot declara.
| Slot | Selector | Obligatorio | Inyección | Sensible | Descripción |
|---|---|---|---|---|---|
| model_pull_token | feature.aggrega.ai/kind=secret | no | env | sí | Optional Hugging Face Hub access token for a gated model. When attached, the weights-pull step receives it as an environment variable, and the server itself never does; unattached, weights are pulled anonymously, which the default model allows. |
Lista de permitidos de autoinyección
No se declararon canales de autoinyección.
Relaciones
Aristas tipadas hacia y desde otras entidades publicadas. Una referencia sin un destino publicado se muestra como una arista rota en lugar de como un enlace.
Usa
Esta entidad no usa nada del catálogo.
Usado por
usa el socket «auth_core»
usa el socket «auth_core»
usa el socket «auth_core»
usa el socket «auth_core»
usa el socket «upstream»
usa el socket «forward_target»
usa el socket «scrape_target»
usa el socket «store»
usa el socket «detector»
usa el socket «detector»
usa el socket «jailbreak_detection»
se aplica a
se aplica a
se aplica a
se aplica a
se aplica a
se aplica a
se aplica a
implementa
implementa