@cap-js/agents 0.9.5 → 0.9.7

package/README.md

@@ -6,138 +6,9 @@
 
 CDS plugin for building agents based on the [A2A](https://a2a-protocol.org) protocol.
 
-## Requirements and Setup
+## Usage
 
-We use the [@capire/bookshop](https://github.com/capire/bookshop) as a running sample hereinafter. Clone it and open it in VSCode as follows:
-
-```bash
-git clone https://github.com/capire/bookshop
-code bookshop
-```
-
-Within your project root run this to add the plugin:
-
-```bash
-npm add @cap-js/agents
-```
-
-Annotate the `CatalogService` with `@agent`:
-
-```cds
-// srv/cat-service.cds
-...
-annotate CatalogService with @agent;
-```
-
-Start your server with `cds watch` and note that the A2A protocol gets served:
-
-```bash
-[cds] - serving CatalogService {
-  at: [ ..., '/a2a/browse' ],
-  ...
-}
-```
-
-For local development, the plugin serves a chat preview at `http://localhost:4004/a2a/browse/preview/` and features a mock LLM.
-To use a real LLM, simply bind your app to an existing instance of SAP AI Core and start your server in `hybrid` profile:
-
-```bash
-cds bind -2 <instance>
-cds w --profile hybrid
-```
-
-See [SAP AI Core → Create a Service Instance](https://help.sap.com/docs/sap-ai-core/sap-ai-core-service-guide/create-service-instance) for how to create an instance.
-
-## Ways to Build Agents
-
-Both approaches start from a CDS service annotated with `@agent` (as shown in [Requirements and Setup](#requirements-and-setup)). They differ only in how the agent's behaviour is defined.
-
-### Agentify Existing CAP Services
-
-With `@agent` alone, the plugin auto-generates tools from the service's entities and actions, creates a [ReAct agent](https://arxiv.org/abs/2210.03629) loop, and serves it as a remote agent — no code required.
-
-```cds
-@agent
-service CatalogService {
-  entity Books as projection on my.Books;
-  action submitOrder(book: Books:ID, quantity: Integer) returns { stock: Integer };
-}
-```
-
-### Markdown-Based Agents
-
-To define an agent's identity, behaviour, and skills explicitly, add a sibling directory matching the slugified service name. When present, it replaces the default agentification: instead of the auto-generated ReAct agent, the plugin auto-builds the agent from the directory at startup — no JavaScript handler required.
-
-```
-srv/
-├─ cat-service.cds
-└─ catalog-agent/                ← matches the slugified service name
-   ├─ AGENTS.md                  ← agent identity + behaviour
-   └─ skills/
-      └─ book-purchase/
-         └─ SKILL.md             ← workflow + examples
-```
-
-`AGENTS.md` defines who the agent is. The frontmatter populates the agent card;
-the body is the agent's system prompt:
-
-```md
----
-name: catalog-agent
-version: "1.0.0"
-description: >
-  Bookshop assistant for placing book orders on behalf of the user.
----
-
-# Catalog Agent
-
-## Identity
-
-You are the **Catalog Agent**, a helpful assistant for the capire bookshop.
-...
-```
-
-## Human-in-the-Loop
-
-Annotate a CDS action with `@agent.hitl` to require human approval before the agent may execute it. When the agent decides to call the action, the task pauses and transitions to the A2A [`input-required`](https://a2a-protocol.org/latest/specification/#413-taskstate) state instead of running the action immediately.
-
-```cds
-// srv/cat-service.cds
-...
-annotate CatalogService.submitOrder with @agent.hitl;
-```
-
-## Configuration
-
-The LLM used by an agent is configured via `cds.requires.llm`. You can provide a `kind` as with [any required service](https://cap.cloud.sap/docs/node.js/core-services#required-services).
-
-```jsonc
-"cds": {
-  "requires": {
-    "llm": {
-      "kind": "aicore",
-      "model": "anthropic--claude-4.6-sonnet"
-    }
-  }
-}
-```
-
-| Kind     | Description                                                           |
-| -------- | --------------------------------------------------------------------- |
-| `aicore` | The default for `production` and `hybrid`, connects to SAP AI Core    |
-| `mock`   | The default for `development`, provides a mocked response when called |
-
-## Advanced
-
-The following capabilities are experimental and documented separately. Their public surface may change.
-
-- [Connectivity](.docs/connectivity.md) — destination-based connectivity, `AICORE_SERVICE_KEY` / `ANTHROPIC_API_KEY`, and the `anthropic` kind
-- [Configuration](.docs/configuration.md) — using multiple models, global and per-service settings, file I/O, and push notifications
-- [Quota Enforcement](.docs/quota.md) — configurable rate limits and resource quotas
-- [Audit Logging](.docs/audit-logging.md) — immutable audit trail of agent decisions and tool usage
-- [Data Privacy](.docs/data-privacy.md) — deletion of message history
-- [Telemetry](.docs/telemetry.md) — OpenTelemetry metrics, tracing, and MLflow export
-- [Content Filter](.docs/content-filter.md) — SAP AI Core content filtering and prompt injection shielding
+For detailed instructions about setup and usage, refer to the [official documentation](https://cap.cloud.sap/docs/guides/ai/cap-agents).
 
 ## Support, Feedback, Contributing
 

package/_i18n/messages_ar.properties

@@ -0,0 +1,48 @@
+# Agent status messages (shown during task execution)
+#XMSG Status info message send to the user client when a tool is called. {0} = tool name
+agent_status_calling_tools=جارٍ استدعاء {0}
+#XMSG Status info message send to the user client when an entity is being queried. {0} = entity name
+agent_status_querying=جارٍ الاستعلام عن {0}
+#XMSG Status info message send to the user client after a tool result returned and the agent is processing it
+agent_status_processing_response=استجابة أداة المعالجة
+#XMSG Status info message send to the user client after multiple tool results returned and the agent is processing it
+agent_status_processing_responses=استجابات أداة المعالجة
+agent_status_summarizing_progress=تمت مقاطعة الوكيل. جارٍ تلخيص التقدم.
+AGENT_SUMMARY_QUOTA_FALLBACK=تم إيقاف المهمة لأنها وصلت إلى حد استهلاك المورد الخاص بها. ربما تم تنفيذ بعض الأعمال - يرجى فحص النتائج أو إعادة المحاولة باستخدام طلب أبسط.
+AGENT_SUMMARY_TIMEOUT_FALLBACK=تم إيقاف المهمة لأنها وصلت إلى الحد الزمني المحدد لها. قد يكون التقدم غير مكتمل. هل تريد متابعة التشغيل أو الإيقاف؟
+
+# Authentication
+#XMSG Error message for HTTP 401 status
+UNAUTHORIZED=غير مفوَّض
+#XMSG Error message for HTTP 403 status
+FORBIDDEN=ممنوع
+
+# Input validation
+#XMSG Error message send to the users client when their message exceeds the limit. {0} = number of allowed characters
+MESSAGE_TOO_LONG=يجب ألا تتجاوز الرسالة {0} من الحروف.
+
+# Quota enforcement
+QUOTA_CONCURRENT_TASKS=تم الوصول إلى الحد الأقصى البالغ {0} من المهام المتزامنة. يرجى المحاولة مرة أخرى لاحقًا.
+QUOTA_CONCURRENT_TASKS_PER_USER=تم الوصول إلى الحد الأقصى البالغ {0} من المهام المتزامنة لكل مستخدم. يرجى إعادة المحاولة لاحقًا.
+QUOTA_TASKS_PER_HOUR=تم الوصول إلى الحد الأقصى البالغ {0} من المهام لكل ساعة لهذه الوحدة المستضافة. يرجى المحاولة مرة أخرى لاحقًا.
+QUOTA_TASKS_PER_HOUR_PER_USER=تم الوصول إلى الحد الأقصى البالغ {0} من المهام لكل ساعة لكل مستخدم. يرجى إعادة المحاولة لاحقًا.
+QUOTA_TOOL_CALLS_PER_HOUR=تم استخدام الحد الأقصى لعدد استدعاءات الأدوات لكل ساعة ({0}). يرجى المحاولة مرة أخرى لاحقًا.
+QUOTA_LLM_TOKENS_PER_DAY=تم استخدام الحد الأقصى لعدد الرموز المميزة لنموذج اللغة الكبير (LLM) لليوم بالفعل ({0}). يرجى المحاولة مرة أخرى غدًا.
+
+# Content filter
+CONTENT_FILTER_BLOCKED=تم إيقاف إدخالك بواسطة عامل تصفية سلامة المحتوى. يرجى إعادة الصياغة.
+CONTENT_FILTER_UNAVAILABLE=فحص سلامة المحتوى غير متوفر مؤقتًا. يرجى المحاولة مرة أخرى.
+CONTENT_FILTER_TOOL_BLOCKED=أسفرت آخر أداة تم استدعاؤها عن نتيجة ضارة. يرجى تجربة اتجاه مختلف لتجنب البيانات الضارة نفسها.
+
+# Push notifications
+PUSH_URL_NOT_ALLOWED=رابط إعادة استدعاء الإشعار المؤقت ''{0}'' غير موجود في المجالات المسموح بها: {1}
+
+# HITL
+HITL_APPROVE=اعتماد
+HITL_REJECT=رفض
+HITL_CONTINUE=متابعة
+HITL_STOP=إيقاف
+RESUME_REQUIRES_TEXT=يجب أن تحتوي رسالة الاستئناف على نص (على سبيل المثال، ''اعتماد'' أو ''رفض'') أو جزء بيانات.
+
+# LLM
+LLM_UNAVAILABLE=الوكيل غير متوفر مؤقتًا. يرجى المحاولة مرة أخرى لاحقًا.

package/_i18n/messages_bg.properties

@@ -0,0 +1,48 @@
+# Agent status messages (shown during task execution)
+#XMSG Status info message send to the user client when a tool is called. {0} = tool name
+agent_status_calling_tools=Извикване на {0}
+#XMSG Status info message send to the user client when an entity is being queried. {0} = entity name
+agent_status_querying=Запитване към {0}
+#XMSG Status info message send to the user client after a tool result returned and the agent is processing it
+agent_status_processing_response=Отговор от инструмент за обработка
+#XMSG Status info message send to the user client after multiple tool results returned and the agent is processing it
+agent_status_processing_responses=Отговори от инструмент за обработка
+agent_status_summarizing_progress=Агентът е прекъснат. Напредъкът се обобщава.
+AGENT_SUMMARY_QUOTA_FALLBACK=Задачата е спряна, тъй като е достигнала лимита си за потребление на ресурси. Част от работата може да е извършена – моля, проверете резултатите или опитайте отново с по-проста заявка.
+AGENT_SUMMARY_TIMEOUT_FALLBACK=Задачата е спряна, тъй като е достигнала времевия си лимит. Напредъкът може да е непълен. Продължаване на изпълнението или спиране?
+
+# Authentication
+#XMSG Error message for HTTP 401 status
+UNAUTHORIZED=Без оторизация
+#XMSG Error message for HTTP 403 status
+FORBIDDEN=Забранено
+
+# Input validation
+#XMSG Error message send to the users client when their message exceeds the limit. {0} = number of allowed characters
+MESSAGE_TOO_LONG=Съобщението не може да надвишава {0} символа.
+
+# Quota enforcement
+QUOTA_CONCURRENT_TASKS=Достигнат е максимумът от {0} едновременни задачи. Моля, опитайте отново по-късно.
+QUOTA_CONCURRENT_TASKS_PER_USER=Достигнат е максимумът от {0} едновременни задачи за един потребител. Моля, опитайте отново по-късно.
+QUOTA_TASKS_PER_HOUR=Достигнат е максимумът от {0} задачи на час за този наемател. Моля, опитайте отново по-късно.
+QUOTA_TASKS_PER_HOUR_PER_USER=Достигнат е максимумът от {0} задачи на час за един потребител. Моля, опитайте отново по-късно.
+QUOTA_TOOL_CALLS_PER_HOUR=Използван е максималният брой извиквания на инструменти на час ({0}). Моля, опитайте отново по-късно.
+QUOTA_LLM_TOKENS_PER_DAY=Максималният брой токени за LLM за днес вече е използван ({0}). Моля, опитайте отново утре.
+
+# Content filter
+CONTENT_FILTER_BLOCKED=Въведените от вас данни са блокирани от филтъра за безопасност на съдържанието. Моля, перифразирайте.
+CONTENT_FILTER_UNAVAILABLE=Проверката за безопасност на съдържанието е временно недостъпна. Моля, опитайте отново.
+CONTENT_FILTER_TOOL_BLOCKED=Последният извикан инструмент върна злонамерен резултат. Моля, опитайте различна посока на действие, за да избегнете същите злонамерени данни.
+
+# Push notifications
+PUSH_URL_NOT_ALLOWED=URL за обратно извикване на насочено известие ''{0}'' не е в разрешените домейни: {1}
+
+# HITL
+HITL_APPROVE=Одобряване
+HITL_REJECT=Отхвърляне
+HITL_CONTINUE=Напред
+HITL_STOP=Стоп
+RESUME_REQUIRES_TEXT=Съобщението за възобновяване трябва да съдържа текст (напр. „одобряване“ или „отхвърляне“) или част с данни.
+
+# LLM
+LLM_UNAVAILABLE=Агентът е временно недостъпен. Моля, опитайте отново по-късно.

package/_i18n/messages_cs.properties

@@ -0,0 +1,48 @@
+# Agent status messages (shown during task execution)
+#XMSG Status info message send to the user client when a tool is called. {0} = tool name
+agent_status_calling_tools=Volající {0}
+#XMSG Status info message send to the user client when an entity is being queried. {0} = entity name
+agent_status_querying=Dotazující {0}
+#XMSG Status info message send to the user client after a tool result returned and the agent is processing it
+agent_status_processing_response=Zpracovává se reakce nástroje
+#XMSG Status info message send to the user client after multiple tool results returned and the agent is processing it
+agent_status_processing_responses=Zpracovávají se reakce nástroje
+agent_status_summarizing_progress=Agent byl přerušen. Shrnuje se pokrok.
+AGENT_SUMMARY_QUOTA_FALLBACK=Úloha byla zastavena, protože dosáhla limitu spotřeby zdrojů. Některé práce byly možná provedeny – zkontrolujte výsledky nebo zkuste to znovu s jednodušším požadavkem.
+AGENT_SUMMARY_TIMEOUT_FALLBACK=Úloha byla zastavena, protože dosáhla svého časového limitu. Pokrok může být neúplný. Pokračovat, nebo zastavit?
+
+# Authentication
+#XMSG Error message for HTTP 401 status
+UNAUTHORIZED=Neautorizováno
+#XMSG Error message for HTTP 403 status
+FORBIDDEN=Zakázáno
+
+# Input validation
+#XMSG Error message send to the users client when their message exceeds the limit. {0} = number of allowed characters
+MESSAGE_TOO_LONG=Zpráva nesmí překročit {0} znaků.
+
+# Quota enforcement
+QUOTA_CONCURRENT_TASKS=Bylo dosaženo maximálního počtu {0} souběžných úloh. Zkuste to znovu později.
+QUOTA_CONCURRENT_TASKS_PER_USER=Bylo dosaženo maximálního počtu {0} souběžných úloh na uživatele. Zkuste to znovu později.
+QUOTA_TASKS_PER_HOUR=Bylo dosaženo maximálního počtu {0} úloh za hodinu pro tohoto tenanta. Zkuste to znovu později.
+QUOTA_TASKS_PER_HOUR_PER_USER=Bylo dosaženo maximálního počtu {0} úloh za hodinu na uživatele. Zkuste to znovu později.
+QUOTA_TOOL_CALLS_PER_HOUR=Bylo použito maximální množství volání nástrojů za hodinu ({0}). Zkuste to znovu později.
+QUOTA_LLM_TOKENS_PER_DAY=Maximální počet tokenů LLM pro dnešek byl již použit ({0}). Zkuste to znovu zítra.
+
+# Content filter
+CONTENT_FILTER_BLOCKED=Váš vstup byl blokován bezpečnostním filtrem obsahu. Změňte prosím formulaci.
+CONTENT_FILTER_UNAVAILABLE=Kontrola bezpečnosti obsahu je dočasně nedostupná. Zkuste to znovu.
+CONTENT_FILTER_TOOL_BLOCKED=Poslední volaný nástroj vrátil škodlivý výsledek. Zkuste jiný směr, abyste předešli stejným škodlivým datům.
+
+# Push notifications
+PUSH_URL_NOT_ALLOWED=URL zpětného volání oznámení push ''{0}'' není povoleno v doménách: {1}
+
+# HITL
+HITL_APPROVE=Schválit
+HITL_REJECT=Zamítnout
+HITL_CONTINUE=Pokračovat
+HITL_STOP=Zastavit
+RESUME_REQUIRES_TEXT=Zpráva o pokračování musí obsahovat text (např. ''schválit'' nebo ''zamítnout'') nebo datovou část.
+
+# LLM
+LLM_UNAVAILABLE=Agent je dočasně nedostupný. Pokuste se prosím znovu později.

package/_i18n/messages_da.properties

@@ -0,0 +1,48 @@
+# Agent status messages (shown during task execution)
+#XMSG Status info message send to the user client when a tool is called. {0} = tool name
+agent_status_calling_tools=Kalder {0}
+#XMSG Status info message send to the user client when an entity is being queried. {0} = entity name
+agent_status_querying=Forespørger om {0}
+#XMSG Status info message send to the user client after a tool result returned and the agent is processing it
+agent_status_processing_response=Behandler svar fra værktøj
+#XMSG Status info message send to the user client after multiple tool results returned and the agent is processing it
+agent_status_processing_responses=Behandler svar fra værktøj
+agent_status_summarizing_progress=Agent blev afbrudt. Opsummerer fremskridt.
+AGENT_SUMMARY_QUOTA_FALLBACK=Opgaven blev stoppet, da den nåede sin ressourceforbrugsgrænse. Noget arbejde kan være udført – kontroller resultaterne, eller prøv igen med en enklere anmodning.
+AGENT_SUMMARY_TIMEOUT_FALLBACK=Opgaven blev stoppet, da den nåede sin tidsgrænse. Fremskridt kan være ufuldstændig. Fortsæt med at køre, eller stop?
+
+# Authentication
+#XMSG Error message for HTTP 401 status
+UNAUTHORIZED=Ikke autoriseret
+#XMSG Error message for HTTP 403 status
+FORBIDDEN=Forbudt
+
+# Input validation
+#XMSG Error message send to the users client when their message exceeds the limit. {0} = number of allowed characters
+MESSAGE_TOO_LONG=Meddelelsen må ikke overskride {0} tegn.
+
+# Quota enforcement
+QUOTA_CONCURRENT_TASKS=Det maksimale antal på {0} samtidige opgaver er nået. Prøv igen senere.
+QUOTA_CONCURRENT_TASKS_PER_USER=Det maksimale antal på {0} samtidige opgaver pr. bruger er nået. Prøv igen senere.
+QUOTA_TASKS_PER_HOUR=Det maksimale antal på {0} opgaver pr. time for denne tenant er nået. Prøv igen senere.
+QUOTA_TASKS_PER_HOUR_PER_USER=Det maksimale antal på {0} opgaver pr. time pr. bruger er nået. Prøv igen senere.
+QUOTA_TOOL_CALLS_PER_HOUR=Den maksimale mængde værktøjskald pr. time blev brugt ({0}). Prøv igen senere.
+QUOTA_LLM_TOKENS_PER_DAY=Det maksimale antal LLM-tokens for i dag er allerede brugt ({0}). Prøv igen i morgen.
+
+# Content filter
+CONTENT_FILTER_BLOCKED=Dit input blev blokeret af filteret for indholdssikkerhed. Vær venlig at omformulere.
+CONTENT_FILTER_UNAVAILABLE=Indholdssikkerhedskontrollen er midlertidigt utilgængelig. Prøv igen.
+CONTENT_FILTER_TOOL_BLOCKED=Det sidste kaldte værktøj gav et skadeligt resultat. Prøv en anden retning for at undgå de samme skadelige data.
+
+# Push notifications
+PUSH_URL_NOT_ALLOWED=Push-meddelelsens callback-URL ''{0}'' er ikke i tilladte domæner: {1}
+
+# HITL
+HITL_APPROVE=Godkend
+HITL_REJECT=Afvis
+HITL_CONTINUE=Fortsæt
+HITL_STOP=Stop
+RESUME_REQUIRES_TEXT=Genoptag meddelelse skal indeholde tekst (f.eks. ''godkend'' eller ''afvis'') eller en datadel.
+
+# LLM
+LLM_UNAVAILABLE=Agenten er midlertidigt utilgængelig. Prøv igen senere.

package/_i18n/messages_de.properties

@@ -0,0 +1,48 @@
+# Agent status messages (shown during task execution)
+#XMSG Status info message send to the user client when a tool is called. {0} = tool name
+agent_status_calling_tools={0} wird aufgerufen
+#XMSG Status info message send to the user client when an entity is being queried. {0} = entity name
+agent_status_querying={0} wird abgefragt
+#XMSG Status info message send to the user client after a tool result returned and the agent is processing it
+agent_status_processing_response=Tool-Antwort wird verarbeitet...
+#XMSG Status info message send to the user client after multiple tool results returned and the agent is processing it
+agent_status_processing_responses=Tool-Antworten werden verarbeitet...
+agent_status_summarizing_progress=Agent unterbrochen. Fortschritt wird zusammengefasst.
+AGENT_SUMMARY_QUOTA_FALLBACK=Die Aufgabe wurde beendet, da das Ressourcen-Limit erreicht wurde. Gegebenenfalls wurden dennoch einige Arbeiten ausgeführt. Überprüfen Sie das Ergebnis oder versuchen Sie es mit einer einfacheren Anforderung erneut.
+AGENT_SUMMARY_TIMEOUT_FALLBACK=Die Aufgabe wurde beendet, da das Zeitlimit erreicht wurde. Der Fortschritt ist gegebenenfalls unvollständig. Mit Ausführung fortfahren oder abbrechen?
+
+# Authentication
+#XMSG Error message for HTTP 401 status
+UNAUTHORIZED=Nicht berechtigt
+#XMSG Error message for HTTP 403 status
+FORBIDDEN=Unzulässig
+
+# Input validation
+#XMSG Error message send to the users client when their message exceeds the limit. {0} = number of allowed characters
+MESSAGE_TOO_LONG=Die Meldung darf {0} Zeichen nicht überschreiten.
+
+# Quota enforcement
+QUOTA_CONCURRENT_TASKS=Maximum von {0} gleichzeitigen Aufgaben erreicht. Versuchen Sie es später erneut.
+QUOTA_CONCURRENT_TASKS_PER_USER=Maximum von {0} gleichzeitigen Aufgaben pro Benutzer erreicht. Versuchen Sie es später erneut.
+QUOTA_TASKS_PER_HOUR=Maximum von {0} gleichzeitigen Aufgaben pro Stunde für diesen Tenant erreicht. Versuchen Sie es später erneut.
+QUOTA_TASKS_PER_HOUR_PER_USER=Maximum von {0} gleichzeitigen Aufgaben pro Stunde erreicht. Versuchen Sie es später erneut.
+QUOTA_TOOL_CALLS_PER_HOUR=Maximale Anzahl an Tool-Aufrufen pro Stunde ({0}) erreicht. Versuchen Sie es später erneut.
+QUOTA_LLM_TOKENS_PER_DAY=Maximale Anzahl an LLM-Tokens für heute ({0}) erreicht. Versuchen Sie es morgen erneut.
+
+# Content filter
+CONTENT_FILTER_BLOCKED=Eingabe wurde durch den Content-Sicherheitsfilter gesperrt. Bitte umformulieren.
+CONTENT_FILTER_UNAVAILABLE=Die Content-Sicherheitsprüfung ist vorübergehend nicht verfügbar. Versuchen Sie es erneut.
+CONTENT_FILTER_TOOL_BLOCKED=Das zuletzt aufgerufene Tool lieferte ein bösartiges Ergebnis. Versuchen Sie etwas anderes, um schädliche Daten zu vermeiden.
+
+# Push notifications
+PUSH_URL_NOT_ALLOWED=Push-Benachrichtigungs-Callback-URL "{0}" ist nicht in der Liste zulässiger Domänen enthalten: {1}
+
+# HITL
+HITL_APPROVE=Genehmigen
+HITL_REJECT=Ablehnen
+HITL_CONTINUE=Weiter
+HITL_STOP=Beenden
+RESUME_REQUIRES_TEXT=Die Resume-Nachricht muss Text (z.B. "genehmigen" oder "ablehnen") oder einen Datenteil enthalten.
+
+# LLM
+LLM_UNAVAILABLE=Der Agent ist vorübergehend nicht verfügbar. Versuchen Sie es später erneut.

package/_i18n/messages_el.properties

@@ -0,0 +1,48 @@
+# Agent status messages (shown during task execution)
+#XMSG Status info message send to the user client when a tool is called. {0} = tool name
+agent_status_calling_tools=Κλήση {0}
+#XMSG Status info message send to the user client when an entity is being queried. {0} = entity name
+agent_status_querying=Γίνεται αναζήτηση {0}
+#XMSG Status info message send to the user client after a tool result returned and the agent is processing it
+agent_status_processing_response=Απάντηση εργαλείου επεξεργασίας
+#XMSG Status info message send to the user client after multiple tool results returned and the agent is processing it
+agent_status_processing_responses=Απαντήσεις εργαλείου επεξεργασίας
+agent_status_summarizing_progress=Ο πράκτορας διακόπηκε. Σύνοψη προόδου.
+AGENT_SUMMARY_QUOTA_FALLBACK=Η εργασία διακόπηκε καθώς έφτασε το όριο ανάλωσης πόρων. Μπορεί να έχει γίνει κάποια εργασία — ελέγξτε τα αποτελέσματα ή δοκιμάστε ξανά με ένα απλούστερο αίτημα.
+AGENT_SUMMARY_TIMEOUT_FALLBACK=Η εργασία διακόπηκε καθώς έφτασε στο χρονικό της όριο. Η πρόοδος μπορεί να είναι ελλιπής. Συνέχεια εκτέλεσης ή διακοπής;
+
+# Authentication
+#XMSG Error message for HTTP 401 status
+UNAUTHORIZED=Μη Εξουσιοδοτημένος
+#XMSG Error message for HTTP 403 status
+FORBIDDEN=Απαγορευμένο
+
+# Input validation
+#XMSG Error message send to the users client when their message exceeds the limit. {0} = number of allowed characters
+MESSAGE_TOO_LONG=Το μήνυμα δεν πρέπει να υπερβαίνει τους {0} χαρακτήρες.
+
+# Quota enforcement
+QUOTA_CONCURRENT_TASKS=Ο μέγιστος αριθμός {0} ταυτόχρονων εργασιών προσεγγίστηκε. Δοκιμάστε ξανά αργότερα.
+QUOTA_CONCURRENT_TASKS_PER_USER=Ο μέγιστος αριθμός {0} ταυτόχρονων εργασιώνανά χρήστη προσεγγίστηκε. Δοκιμάστε ξανά αργότερα.
+QUOTA_TASKS_PER_HOUR=Ο μέγιστος αριθμός {0} εργασιών ανά ώρα για αυτόν τον μισθωτή προσεγγίστηκε. Δοκιμάστε ξανά αργότερα.
+QUOTA_TASKS_PER_HOUR_PER_USER=Ο μέγιστος αριθμός {0} εργασιών ανά χρήστη προσεγγίστηκε. Δοκιμάστε ξανά αργότερα.
+QUOTA_TOOL_CALLS_PER_HOUR=Χρησιμοποιήθηκε ο μέγιστος αριθμός κλήσεων εργαλείων ανά ώρα ({0}). Δοκιμάστε ξανά αργότερα.
+QUOTA_LLM_TOKENS_PER_DAY=Το μέγιστο ποσό διακριτικών LLM για σήμερα χρησιμοποιήθηκε ήδη ({0}). Δοκιμάστε ξανά αύριο.
+
+# Content filter
+CONTENT_FILTER_BLOCKED=Η καταχώρισή σας δεσμεύτηκε από το φίλτρο ασφάλειας περιεχομένου. Αναδιατυπώστε.
+CONTENT_FILTER_UNAVAILABLE=Ο έλεγχος ασφάλειας περιεχομένου είναι προσωρινά μη διαθέσιμος. Δοκιμάστε ξανά.
+CONTENT_FILTER_TOOL_BLOCKED=Το τελευταίο εργαλείο που ονομάζεται απέδωσε ένα κακόβουλο αποτέλεσμα. Δοκιμάστε διαφορετική κατεύθυνση για να αποφύγετε τα ίδια κακόβουλα δεδομένα.
+
+# Push notifications
+PUSH_URL_NOT_ALLOWED=URL ανάκλησης ειδοποίησης push ''{0}'' δεν είναι σε επιτρεπόμενους τομείς: {1}
+
+# HITL
+HITL_APPROVE=Έγκριση
+HITL_REJECT=Απόρριψη
+HITL_CONTINUE=Συνέχεια
+HITL_STOP=Διακοπή
+RESUME_REQUIRES_TEXT=Το μήνυμα συνέχισης πρέπει να περιέχει κείμενο (π.χ. ''έγκριση'' ή ''απόρριψη'') ή μέρος δεδομένων.
+
+# LLM
+LLM_UNAVAILABLE=Ο πράκτορας είναι προσωρινά μη διαθέσιμος. Προσπαθήστε ξανά αργότερα.

package/_i18n/messages_en.properties

@@ -0,0 +1,48 @@
+# Agent status messages (shown during task execution)
+#XMSG Status info message send to the user client when a tool is called. {0} = tool name
+agent_status_calling_tools=Calling {0}
+#XMSG Status info message send to the user client when an entity is being queried. {0} = entity name
+agent_status_querying=Querying {0}
+#XMSG Status info message send to the user client after a tool result returned and the agent is processing it
+agent_status_processing_response=Processing tool response
+#XMSG Status info message send to the user client after multiple tool results returned and the agent is processing it
+agent_status_processing_responses=Processing tool responses
+agent_status_summarizing_progress=Agent interrupted. Summarising progress.
+AGENT_SUMMARY_QUOTA_FALLBACK=The task was stopped as it reached its resource consumption limit. Some work may have been done — please check the results or retry with a simpler request.
+AGENT_SUMMARY_TIMEOUT_FALLBACK=The task was stopped as it reached its time limit. Progress may be incomplete. Continue running or stop?
+
+# Authentication
+#XMSG Error message for HTTP 401 status
+UNAUTHORIZED=Unauthorized
+#XMSG Error message for HTTP 403 status
+FORBIDDEN=Forbidden
+
+# Input validation
+#XMSG Error message send to the users client when their message exceeds the limit. {0} = number of allowed characters
+MESSAGE_TOO_LONG=The message must not exceed {0} characters.
+
+# Quota enforcement
+QUOTA_CONCURRENT_TASKS=The maximum of {0} concurrent tasks is reached. Please try again later.
+QUOTA_CONCURRENT_TASKS_PER_USER=The maximum of {0} concurrent tasks per user is reached. Please try again later.
+QUOTA_TASKS_PER_HOUR=The maximum of {0} tasks per hour for this tenant is reached. Please try again later.
+QUOTA_TASKS_PER_HOUR_PER_USER=The maximum of {0} tasks per hour per user is reached. Please try again later.
+QUOTA_TOOL_CALLS_PER_HOUR=The maximum amount of tool calls per hour was used ({0}). Please try again later.
+QUOTA_LLM_TOKENS_PER_DAY=The maximum amount of LLM tokens for today was already used ({0}). Please try again tomorrow.
+
+# Content filter
+CONTENT_FILTER_BLOCKED=Your input was blocked by the content safety filter. Please rephrase.
+CONTENT_FILTER_UNAVAILABLE=The content safety check is temporarily unavailable. Please try again.
+CONTENT_FILTER_TOOL_BLOCKED=The last tool called yielded a malicious result. Please try a different direction to avoid the same malicious data.
+
+# Push notifications
+PUSH_URL_NOT_ALLOWED=Push notification callback URL ''{0}'' not in allowed domains: {1}
+
+# HITL
+HITL_APPROVE=Approve
+HITL_REJECT=Reject
+HITL_CONTINUE=Continue
+HITL_STOP=Stop
+RESUME_REQUIRES_TEXT=Resume message must contain text (e.g. ''approve'' or ''reject'') or a data part.
+
+# LLM
+LLM_UNAVAILABLE=The agent is temporarily unavailable. Please try again later.

package/_i18n/messages_es.properties

@@ -0,0 +1,48 @@
+# Agent status messages (shown during task execution)
+#XMSG Status info message send to the user client when a tool is called. {0} = tool name
+agent_status_calling_tools=Llamando {0}
+#XMSG Status info message send to the user client when an entity is being queried. {0} = entity name
+agent_status_querying=Consultando {0}
+#XMSG Status info message send to the user client after a tool result returned and the agent is processing it
+agent_status_processing_response=Procesando respuesta de la herramienta
+#XMSG Status info message send to the user client after multiple tool results returned and the agent is processing it
+agent_status_processing_responses=Procesando respuestas de la herramienta
+agent_status_summarizing_progress=Agente interrumpido. Resumiendo el progreso.
+AGENT_SUMMARY_QUOTA_FALLBACK=La tarea se ha detenido porque ha alcanzado su límite de consumo de recursos. Es posible que se haya realizado algún trabajo; verifique los resultados o inténtelo de nuevo con una solicitud más sencilla.
+AGENT_SUMMARY_TIMEOUT_FALLBACK=La tarea se ha detenido porque ha alcanzado su límite de tiempo. Es posible que el progreso esté incompleto. ¿Continuar ejecutándose o detenerse?
+
+# Authentication
+#XMSG Error message for HTTP 401 status
+UNAUTHORIZED=No autorizado
+#XMSG Error message for HTTP 403 status
+FORBIDDEN=Prohibido
+
+# Input validation
+#XMSG Error message send to the users client when their message exceeds the limit. {0} = number of allowed characters
+MESSAGE_TOO_LONG=El mensaje no debe superar los {0} caracteres.
+
+# Quota enforcement
+QUOTA_CONCURRENT_TASKS=Se ha alcanzado el máximo de {0} tareas simultáneas. Inténtelo de nuevo más tarde.
+QUOTA_CONCURRENT_TASKS_PER_USER=Se ha alcanzado el máximo de {0} tareas simultáneas por usuario. Inténtelo de nuevo más tarde.
+QUOTA_TASKS_PER_HOUR=Se ha alcanzado el máximo de {0} tareas por hora para este arrendatario. Inténtelo de nuevo más tarde.
+QUOTA_TASKS_PER_HOUR_PER_USER=Se ha alcanzado el máximo de {0} tareas por hora por usuario. Inténtelo de nuevo más tarde.
+QUOTA_TOOL_CALLS_PER_HOUR=Se ha utilizado la cantidad máxima de llamadas de herramienta por hora ({0}). Inténtelo de nuevo más tarde.
+QUOTA_LLM_TOKENS_PER_DAY=La cantidad máxima de tokens LLM para hoy ya se ha utilizado ({0}). Inténtelo de nuevo mañana.
+
+# Content filter
+CONTENT_FILTER_BLOCKED=Su entrada ha sido bloqueada por el filtro de seguridad de contenido. Reformule la frase.
+CONTENT_FILTER_UNAVAILABLE=La verificación de seguridad del contenido no está disponible temporalmente. Inténtelo de nuevo.
+CONTENT_FILTER_TOOL_BLOCKED=La última herramienta llamada a devuelto un resultado malicioso. Intente con una dirección diferente para evitar los mismos datos maliciosos.
+
+# Push notifications
+PUSH_URL_NOT_ALLOWED=La URL de llamada de retorno de notificación push ''{0}'' no está en los dominios permitidos: {1}
+
+# HITL
+HITL_APPROVE=Aprobar
+HITL_REJECT=Rechazar
+HITL_CONTINUE=Continuar
+HITL_STOP=Detener
+RESUME_REQUIRES_TEXT=El mensaje de reanudación debe contener texto (p. ej. ''autorizar'' o ''rechazar'') o una parte de datos.
+
+# LLM
+LLM_UNAVAILABLE=El agente no está disponible temporalmente. Inténtelo de nuevo más tarde.

package/_i18n/messages_es_MX.properties

@@ -0,0 +1,48 @@
+# Agent status messages (shown during task execution)
+#XMSG Status info message send to the user client when a tool is called. {0} = tool name
+agent_status_calling_tools=Llamando {0}
+#XMSG Status info message send to the user client when an entity is being queried. {0} = entity name
+agent_status_querying=Consultando {0}
+#XMSG Status info message send to the user client after a tool result returned and the agent is processing it
+agent_status_processing_response=Procesando respuesta de la herramienta
+#XMSG Status info message send to the user client after multiple tool results returned and the agent is processing it
+agent_status_processing_responses=Procesando respuestas de la herramienta
+agent_status_summarizing_progress=Agente interrumpido. Resumiendo el progreso.
+AGENT_SUMMARY_QUOTA_FALLBACK=La tarea se ha detenido porque ha alcanzado su límite de consumo de recursos. Es posible que se haya realizado algún trabajo; verifique los resultados o inténtelo de nuevo con una solicitud más sencilla.
+AGENT_SUMMARY_TIMEOUT_FALLBACK=La tarea se ha detenido porque ha alcanzado su límite de tiempo. Es posible que el progreso esté incompleto. ¿Continuar ejecutándose o detenerse?
+
+# Authentication
+#XMSG Error message for HTTP 401 status
+UNAUTHORIZED=No autorizado
+#XMSG Error message for HTTP 403 status
+FORBIDDEN=Prohibido
+
+# Input validation
+#XMSG Error message send to the users client when their message exceeds the limit. {0} = number of allowed characters
+MESSAGE_TOO_LONG=El mensaje no debe superar los {0} caracteres.
+
+# Quota enforcement
+QUOTA_CONCURRENT_TASKS=Se ha alcanzado el máximo de {0} tareas simultáneas. Inténtelo de nuevo más tarde.
+QUOTA_CONCURRENT_TASKS_PER_USER=Se ha alcanzado el máximo de {0} tareas simultáneas por usuario. Inténtelo de nuevo más tarde.
+QUOTA_TASKS_PER_HOUR=Se ha alcanzado el máximo de {0} tareas por hora para este arrendatario. Inténtelo de nuevo más tarde.
+QUOTA_TASKS_PER_HOUR_PER_USER=Se ha alcanzado el máximo de {0} tareas por hora por usuario. Inténtelo de nuevo más tarde.
+QUOTA_TOOL_CALLS_PER_HOUR=Se ha utilizado la cantidad máxima de llamadas de herramienta por hora ({0}). Inténtelo de nuevo más tarde.
+QUOTA_LLM_TOKENS_PER_DAY=La cantidad máxima de tokens LLM para hoy ya se ha utilizado ({0}). Inténtelo de nuevo mañana.
+
+# Content filter
+CONTENT_FILTER_BLOCKED=Su entrada ha sido bloqueada por el filtro de seguridad de contenido. Reformule la frase.
+CONTENT_FILTER_UNAVAILABLE=La verificación de seguridad del contenido no está disponible temporalmente. Inténtelo de nuevo.
+CONTENT_FILTER_TOOL_BLOCKED=La última herramienta llamada a devuelto un resultado malicioso. Intente con una dirección diferente para evitar los mismos datos maliciosos.
+
+# Push notifications
+PUSH_URL_NOT_ALLOWED=La URL de llamada de retorno de notificación push ''{0}'' no está en los dominios permitidos: {1}
+
+# HITL
+HITL_APPROVE=Aprobar
+HITL_REJECT=Rechazar
+HITL_CONTINUE=Continuar
+HITL_STOP=Detener
+RESUME_REQUIRES_TEXT=El mensaje de reanudación debe contener texto (p. ej. ''autorizar'' o ''rechazar'') o una parte de datos.
+
+# LLM
+LLM_UNAVAILABLE=El agente no está disponible temporalmente. Inténtelo de nuevo más tarde.

package/_i18n/messages_fi.properties

@@ -0,0 +1,48 @@
+# Agent status messages (shown during task execution)
+#XMSG Status info message send to the user client when a tool is called. {0} = tool name
+agent_status_calling_tools=Kutsutaan {0}
+#XMSG Status info message send to the user client when an entity is being queried. {0} = entity name
+agent_status_querying=Kysellään kohdetta {0}
+#XMSG Status info message send to the user client after a tool result returned and the agent is processing it
+agent_status_processing_response=Käsittelytyökalun vastaus
+#XMSG Status info message send to the user client after multiple tool results returned and the agent is processing it
+agent_status_processing_responses=Käsittelytyökalun vastaukset
+agent_status_summarizing_progress=Agentti keskeytetty. Tehdään yhteenvetoa edistymisestä.
+AGENT_SUMMARY_QUOTA_FALLBACK=Tehtävä pysäytettiin, koska sen resurssikulutuksen raja saavutettiin. Joitakin töitä on saatettu tehdä — tarkista tulokset tai yritä uudelleen yksinkertaisemmalla pyynnöllä.
+AGENT_SUMMARY_TIMEOUT_FALLBACK=Tehtävä pysäytettiin, koska se saavutti aikarajoituksensa. Edistyminen voi olla puutteellista. Jatketaanko suoritusta vai pysäytetäänkö?
+
+# Authentication
+#XMSG Error message for HTTP 401 status
+UNAUTHORIZED=Ei-oikeutettu
+#XMSG Error message for HTTP 403 status
+FORBIDDEN=Kielletty
+
+# Input validation
+#XMSG Error message send to the users client when their message exceeds the limit. {0} = number of allowed characters
+MESSAGE_TOO_LONG=Viesti ei saa ylittää {0} merkkiä.
+
+# Quota enforcement
+QUOTA_CONCURRENT_TASKS=Rinnakkaisten tehtävien {0} maksimimäärä on saavutettu. Yritä myöhemmin uudelleen.
+QUOTA_CONCURRENT_TASKS_PER_USER=Rinnakkaisten tehtävien {0} maksimimäärä käyttäjää kohti on saavutettu. Yritä myöhemmin uudelleen.
+QUOTA_TASKS_PER_HOUR=Tämän vuokralaisen rinnakkaisten tehtävien {0} maksimimäärä on saavutettu. Yritä myöhemmin uudelleen.
+QUOTA_TASKS_PER_HOUR_PER_USER=Tehtävien {0} maksimimäärä käyttäjää kohti on saavutettu. Yritä myöhemmin uudelleen.
+QUOTA_TOOL_CALLS_PER_HOUR=Työkalukutsujen tuntikohtainen maksimimäärä on käytetty ({0}). Yritä myöhemmin uudelleen.
+QUOTA_LLM_TOKENS_PER_DAY=Tämän päivän LLM-tunnisteiden maksimimäärä on jo käytetty ({0}). Yritä uudelleen huomenna.
+
+# Content filter
+CONTENT_FILTER_BLOCKED=Sisällön turvallisuussuodatin lukitsi syöttämäsi tiedot. Muotoile uudelleen.
+CONTENT_FILTER_UNAVAILABLE=Sisällön turvallisuustarkistus on tilapäisesti poissa käytöstä. Yritä uudelleen.
+CONTENT_FILTER_TOOL_BLOCKED=Viimeksi kutsuttu työkalu tuotti haitallista tietoa. Yritä eri suuntaa välttääksesi samat haitalliset tiedot.
+
+# Push notifications
+PUSH_URL_NOT_ALLOWED=Push-ilmoituksen takaisinkutsun URL "{0}" ei ole sallituilla alueilla: {1}
+
+# HITL
+HITL_APPROVE=Hyväksy
+HITL_REJECT=Hylkää
+HITL_CONTINUE=Jatka
+HITL_STOP=Pysäytä
+RESUME_REQUIRES_TEXT=Jatkamista koskevan viestin on sisällettävä tekstiä (esim. "hyväksy" tai "hylkää") tai tieto-osa.
+
+# LLM
+LLM_UNAVAILABLE=Agentti on väliaikaisesti pois käytöstä. Yritä myöhemmin uudelleen.