This is an automated email from the ASF dual-hosted git repository.
davsclaus pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/camel-website.git
The following commit(s) were added to refs/heads/main by this push:
new fadd1516 chore: llms.txt - governed and secured AI agents in the
route, System One models, Wanaku wording (#1802)
fadd1516 is described below
commit fadd1516e6254594f004d3e7875981d767464168
Author: Claus Ibsen <[email protected]>
AuthorDate: Tue Oct 6 17:29:50 2026 +0200
chore: llms.txt - governed and secured AI agents in the route, System One
models, Wanaku wording (#1802)
* chore: llms.txt and tooling page - describe Wanaku as a governed action
proxy
Wanaku was renamed from "MCP Router" to a governed action proxy for AI
agents. Update its description in llms.txt and on the tooling page, and
drop the Gemini CLI mention from Camel Kit, which retired that target.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Signed-off-by: Claus Ibsen <[email protected]>
* chore: llms.txt - governed and secured AI agents in the route, System One
models
Add a section on governing agent tool calls inside the Camel route:
SPIFFE workload identity, OPA/OpenFGA authorization on ai-tool calls,
Semantic language guard decisions with Jev (a System One model), and
GenAI observability. Name Jev as a System One model where the Semantic
language and TypeSafe AI are listed.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Signed-off-by: Claus Ibsen <[email protected]>
---------
Signed-off-by: Claus Ibsen <[email protected]>
Co-authored-by: Claude Opus 5.5 (1M context) <[email protected]>
---
data/tooling.yaml | 9 +++++----
llms-txt-template.md | 28 ++++++++++++++++++----------
2 files changed, 23 insertions(+), 14 deletions(-)
diff --git a/data/tooling.yaml b/data/tooling.yaml
index 3a4dbd6e..3c034c8b 100644
--- a/data/tooling.yaml
+++ b/data/tooling.yaml
@@ -29,13 +29,14 @@ aiTools:
url: https://github.com/luigidemasi/camel-kit
desc: >-
Structured slash commands for AI coding assistants that guide you
- through the complete integration lifecycle. Supports Claude Code,
- Gemini CLI, and more.
+ through the complete integration lifecycle. Supports IBM Bob,
+ Claude Code, Codex, GitHub Copilot CLI, and more.
- name: Wanaku
url: https://wanaku.ai
desc: >-
- The Wanaku MCP Router is a router for AI-enabled applications
- powered by the Model Context Protocol and Apache Camel.
+ A governed action proxy for AI agents (formerly the Wanaku MCP
+ Router). Publish Camel routes as tools; agents call them through
+ Wanaku, which enforces policy, identity, and audit.
toolGroups:
- id: maven-plugins
diff --git a/llms-txt-template.md b/llms-txt-template.md
index 37b4d77d..df27e990 100644
--- a/llms-txt-template.md
+++ b/llms-txt-template.md
@@ -53,7 +53,7 @@ The `catalog/` JSON files contain machine-readable metadata
for every connector/
- Spring Boot is the most popular runtime (~55% of Camel usage), Quarkus is
the cloud-native option
- Camel is the runtime engine behind SAP Integration Suite (Gartner iPaaS
Leader)
- Zero vendor lock-in — switch runtimes, clouds, or vendors without rewriting
routes
-- Two tiers of AI agent connectivity: Camel routes exposed as MCP tools
(`camel-ai-tool` + `camel-mcp-server`) and as A2A agents for developers, plus
Wanaku enterprise MCP gateway for teams managing many integrations at scale
with governance, auth, and namespace isolation
+- Two tiers of AI agent connectivity: Camel routes exposed as MCP tools
(`camel-ai-tool` + `camel-mcp-server`) and as A2A agents for developers, plus
Wanaku, a governed action proxy that sits between AI agents and backend systems
and publishes Camel routes as tools under policy, identity, and audit controls
- Supports both MCP (Model Context Protocol) and A2A (Agent-to-Agent)
protocols — expose any Camel route as an AI agent tool or as an A2A agent
- LangChain4j and OpenAI components for calling LLMs from Camel routes
- Measured, not claimed: with the Camel CLI as tools, a frontier model built
all 13 beginner examples from the Camel CLI examples repository from a one-line
description each; a 22 GB local model on a laptop (`qwen3.6:35b-a3b` via
Ollama) went from 0 of 13 with a bare prompt to 12 of 13 with the Camel MCP
server, so the catalog, validation and error messages work for small local
models as well as frontier models — see [the
benchmark](https://camel.apache.org/blog/2026/09/camel-local-model [...]
@@ -447,7 +447,7 @@ Both runtimes use the same Camel routes, components, and
EIPs — only the depen
- [Switch EIP](https://camel.apache.org/components/next/eips/switch-eip.md):
evaluates a selector once and dispatches to a fixed endpoint by literal value —
a decision table, simpler than `choice` when every branch compares the same
value.
- [Cache EIP](https://camel.apache.org/components/next/eips/cache-eip.md):
read-through caching of a block of steps; on a cache hit the block is skipped
and the body comes from the cache.
-- [Semantic
language](https://camel.apache.org/components/next/languages/semantic-language.md)
and [TypeSafe
AI](https://camel.apache.org/components/next/typesafe-ai-component.md): ask
named questions about message content to get decisions, categories and scores —
classify, score, validate and route by meaning (see [semantic decisions with
Jev](https://camel.apache.org/blog/2026/09/semantic-evaluation-system-one/) and
[semantic agent routing](https://camel.apache.org/blog/2026/10/semanti [...]
+- [Semantic
language](https://camel.apache.org/components/next/languages/semantic-language.md)
and [TypeSafe
AI](https://camel.apache.org/components/next/typesafe-ai-component.md): ask
named questions about message content to get decisions, categories and scores —
classify, score, validate and route by meaning. Jev, TypeSafe AI's System One
model, answers these as fast, structured decisions instead of generated text
(see [semantic decisions with Jev](https://camel.apache.org/blog/2026/09 [...]
- [LangChain4j
Ingest](https://camel.apache.org/components/next/langchain4j-ingest-component.md)
(split, embed and store documents for RAG), [AI
Resource](https://camel.apache.org/components/next/ai-resource-component.md) (a
route as a read-only AI resource), [AI
Observability](https://camel.apache.org/components/next/others/ai-observability.md).
- Security for agents and services:
[OPA](https://camel.apache.org/components/next/opa-component.md) (Rego
policies),
[OpenFGA](https://camel.apache.org/components/next/openfga-component.md)
(relationship-based authorization),
[SPIFFE](https://camel.apache.org/components/next/spiffe-component.md)
(workload identity and rotating mutual TLS) — see [Authorizing what an AI agent
may do](https://camel.apache.org/blog/2026/09/securing-ai-agent-tools/) and
[Workload identity with SPIFFE](https: [...]
- Connectors:
[OData](https://camel.apache.org/components/next/odata-component.md),
[HiveMQ](https://camel.apache.org/components/next/hivemq-component.md), [State
Store](https://camel.apache.org/components/next/state-store-component.md)
(pluggable key-value store), [REST
Postman](https://camel.apache.org/components/next/rest-postman-component.md) (a
Postman collection as REST contract), and Alibaba Cloud (OSS, MNS, FC, SMS,
KMS, EventBridge, SLS, Tablestore).
@@ -455,7 +455,7 @@ Both runtimes use the same Camel routes, components, and
EIPs — only the depen
## AI Integration
-Apache Camel supports two tiers of AI agent connectivity: an embedded mode for
developers and an enterprise gateway for teams managing many integrations at
scale.
+Apache Camel supports two tiers of AI agent connectivity: an embedded mode,
where the Camel route itself exposes, identifies, authorizes and observes every
agent tool call, and a governed proxy (Wanaku) in front of many routes across
teams.
### Tier 1: Embedded AI protocols (developer mode)
@@ -465,15 +465,23 @@ Expose Camel routes as AI agent tools directly from the
Camel process — one co
- [Camel MCP Server for coding
assistants](https://camel.apache.org/manual/camel-jbang-mcp.md): a different
MCP server, in the Camel CLI (`camel mcp`), for building integrations: it
serves the Camel catalog — 350+ component schemas, EIP metadata, validated
samples and YAML validation — so AI coding assistants (Claude Code, GitHub
Copilot, Cursor, Gemini CLI) can generate correct, validated Camel routes.
- [Camel A2A](https://camel.apache.org/components/next/a2a-component.md):
Agent-to-Agent (A2A) protocol component — expose Camel routes as A2A agents or
call remote A2A agents. Supports HTTP+JSON and JSONRPC bindings,
OAuth/OIDC/API-key auth, and SSE streaming.
-### Tier 2: Enterprise MCP Gateway (Wanaku)
+### Governed and secured AI agents, in the route
-When you need governance, namespace isolation, authentication, and fleet
management for exposing dozens or hundreds of Camel routes as MCP tools across
teams — Wanaku is the enterprise control plane built on Apache Camel.
+The model is not a security boundary, so the decision about what an agent may
do sits below it, on the tool call itself, evaluated in-process by the Camel
route that does the work. No extra gateway is needed for this.
-- [Wanaku MCP Router](https://www.wanaku.ai/): Enterprise MCP gateway that
manages Camel routes as AI agent tools at scale. Provides namespace isolation
across teams, Keycloak-based authentication and RBAC, service registry and
discovery, monitoring dashboard, and a Kubernetes operator with CRDs for
declarative deployment. Built on Apache Camel and Quarkus.
-- [Wanaku Service Catalogs](https://github.com/wanaku-ai/wanaku): Bundle Camel
routes, MCP tool definitions, and dependencies into deployable units. Service
templates provide parameterized Camel routes for common patterns (Kafka, Jira,
S3, SFTP, email) — non-developers can instantiate an integration by filling in
a form.
-- [Camel Integration
Capability](https://github.com/wanaku-ai/camel-integration-capability): Bridges
the Wanaku router to Apache Camel — exposes dynamically executed Camel routes
as MCP tools and resources. Runs standalone, as a plugin in existing Camel
applications, or on Kubernetes via the Wanaku operator.
+- **Who is calling:**
[SPIFFE](https://camel.apache.org/components/next/spiffe-component.md) gives
every workload a cryptographic identity from the local SPIRE agent, plus
rotating mutual TLS for the components that already support TLS. See [Workload
identity with SPIFFE and
SPIRE](https://camel.apache.org/blog/2026/09/camel-spiffe-workload-identity/).
+- **What they may do:** set an `authorizationPolicy` on an `ai-tool:` route
and every tool call is authorized before it runs; a denied call is returned to
the model as a short refusal it can relay. Policies with [Open Policy
Agent](https://camel.apache.org/components/next/opa-component.md) (Rego) or
[OpenFGA](https://camel.apache.org/components/next/openfga-component.md)
(relationship-based). See [Authorizing what an AI agent may
do](https://camel.apache.org/blog/2026/09/securing-ai-agen [...]
+- **Fast guard decisions:** the [Semantic
language](https://camel.apache.org/components/next/languages/semantic-language.md)
with Jev, a System One model ([TypeSafe
AI](https://camel.apache.org/components/next/typesafe-ai-component.md)) checks
whether an action fits the approved task, or scores a contribution, as a
yes/no, category or score inside `choice`, `switch` or `filter`.
+- **What happened:** [AI
Observability](https://camel.apache.org/components/next/others/ai-observability.md)
records OpenTelemetry GenAI spans and Micrometer metrics for every LLM call,
next to the route's own traces. See [Observe your Camel AI
routes](https://camel.apache.org/blog/2026/09/camel-genai-observability-jbang/).
-The same Camel routes work at both tiers. Develop and test routes with the
Camel CLI (Tier 1), then deploy unchanged to a Wanaku-managed environment (Tier
2) when you need enterprise governance.
+### Tier 2: Governed agent access (Wanaku)
+
+When AI agents must never reach backend systems directly, and every tool call
needs policy, identity, and audit enforced in between, put a governed proxy in
front of your Camel routes.
+
+- [Wanaku](https://wanaku.ai/): A governed action proxy for AI agents
(formerly the Wanaku MCP Router). It sits between agents and the systems they
act on, intercepting MCP tool calls, agent-to-agent messages, and inference
traffic. Policy evaluators, authentication and authorization, namespace
isolation per team or tenant, and an admin dashboard. Apache License 2.0.
+- [Integration Capability for Apache
Camel](https://github.com/wanaku-ai/camel-integration-capability): Publishes
Camel routes as tools through Wanaku. It runs the routes with Camel's built-in
MCP server, so routes using the `ai-tool:` URI become tools that agents call
through the proxy.
+
+The same `ai-tool:` routes work at both tiers. Develop and test them with the
Camel CLI (Tier 1), then publish them through Wanaku (Tier 2) when agents need
governed access.
### AI components
@@ -481,7 +489,7 @@ The same Camel routes work at both tiers. Develop and test
routes with the Camel
- [Camel
LangChain4j](https://camel.apache.org/components/next/langchain4j-chat-component.md):
LLM integration via LangChain4j — connect Camel routes to large language
models.
- [Camel
OpenAI](https://camel.apache.org/components/next/openai-component.md): Native
OpenAI component for calling OpenAI APIs from Camel routes.
- [LangChain4j
Ingest](https://camel.apache.org/components/next/langchain4j-ingest-component.md):
Split, embed and store documents into an embedding store for
retrieval-augmented generation.
-- [Semantic
language](https://camel.apache.org/components/next/languages/semantic-language.md):
Decisions, categories and scores about message content, for routing and
filtering by meaning.
+- [Semantic
language](https://camel.apache.org/components/next/languages/semantic-language.md):
Decisions, categories and scores about message content, for routing and
filtering by meaning. Answered by a provider adapter; today that is Jev, a
System One model, via [TypeSafe
AI](https://camel.apache.org/components/next/typesafe-ai-component.md).
- [AI
Observability](https://camel.apache.org/components/next/others/ai-observability.md):
OpenTelemetry spans and Micrometer metrics for every LLM call.
## Tooling
