docs: add camel-karaf security model, SECURITY.md and AGENTS.md

[oscerd]

What

Adds a Karaf-specific threat model for camel-karaf, plus the supporting
SECURITY.md reporting pointer and AGENTS.md AI-agent guidelines — mirroring
the house style of apache/camel
(docs/user-manual/.../security-model.adoc + root SECURITY.md + AGENTS.md
with a Security Model section + CLAUDE.md symlink).

File
docs/modules/ROOT/pages/security-model.adoc	the threat model (new)
docs/modules/ROOT/nav.adoc	links the new page in the user guide nav
SECURITY.md	GitHub-rendered reporting pointer to the model + Apache Camel private process (new)
AGENTS.md (+ CLAUDE.md symlink)	AI-agent guidelines incl. a Security Model summary (new)

Why / approach

camel-karaf is a runtime adapter, not a new data plane. The data-plane
vulnerability classes (unsafe deserialization, XXE, expression/template
injection, path traversal, SSRF, header/bean-dispatch abuse, auth bypass,
information disclosure, insecure defaults, back-end query injection) live in
Apache Camel core and components. The model therefore defers those to the
canonical Apache Camel Security Model
and documents only the Karaf delta:

• the camel-core-osgi resolution layer (must not widen, for untrusted
  message data, a class/bean/component sink flat-classpath Camel core had closed);
• the Blueprint DSL as trusted route-author code (equivalent to Java/XML/YAML DSL);
• the Karaf shell as a management surface (analogous to camel-management/JMX);
• the feature / Pax-URL install model as trusted operator action;
• OSGi is not a security sandbox — the headline false-friend.

It includes a component-family map, input-trust matrix, adversary model,
properties provided / not provided, known non-findings (for automated triage),
a closed set of triage dispositions, and a dated ratification record. The model
is fully ratified; provenance tags are retained as the chain of authority
cited when closing reports.

Produced with the threat-model-producer rubric. No code changes; docs only.
Build gate not affected.

---

Claude Code on behalf of Andrea Cosentino

🤖 Generated with Claude Code

[jbonofre]

The ASF header is missing here.

[oscerd]

Added commit 627dc0ce7 converting the three in-site link:https://camel.apache.org/... URLs in security-model.adoc to root-relative link:/security/ form. This matches apache/camel core security-model.adoc (apache/camel#23224) and prevents this PR from introducing camel/relative-links failures in the camel-website build once merged (same fix shipped for the sibling per-subproject models: apache/camel-kamelets#2836, apache/camel-kafka-connector#1774, apache/camel-quarkus#8661). Documentation-only; link targets unchanged.

Claude Code on behalf of Andrea Cosentino

[oscerd]

@jbonofre addressed in 605b93c3c: added the ASF license header to AGENTS.md, SECURITY.md (HTML-comment form, matching README.md) and docs/modules/ROOT/pages/security-model.adoc (AsciiDoc //// form). CLAUDE.md is a symlink to AGENTS.md, so it inherits the header transitively — kept it as a symlink rather than converting it to a regular file. Re-requesting review.

Claude Code on behalf of Andrea Cosentino