fix(ontology): assign CLIENT role to message producer types (#253)

Types injecting KafkaTemplate, RabbitTemplate, JmsTemplate,
StreamBridge, or ApplicationEventPublisher now get CLIENT role
(symmetry with CONTROLLER covering both HTTP and messaging inbound).
Previously these fell through to OTHER despite having MESSAGE_PRODUCER
capability. Ontology bumped 15 → 16; re-index required.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Author: HumanBean17

agents/java-codebase-rag-explorer.md

@@ -17,7 +17,7 @@ You are a codebase navigation specialist powered by the java-codebase-rag MCP.
 
 Java production sources plus SQL and YAML (use `search` `table`: `java`, `sql`, `yaml`, or `all`).
 
-## Ontology: 15
+## Ontology: 16
 
 If results look structurally wrong or empty across tools, the index may be missing, stale, or built with a different `ontology_version`; you cannot re-index via MCP — ask the operator to rebuild.
 
@@ -243,7 +243,7 @@ Returns **edges** with `attrs` (`confidence`, `strategy`, `match`, … on cross-
 | `COMPONENT` | General Spring component |
 | `CONFIG` | `@Configuration` class |
 | `ENTITY` | JPA / persistence entity |
-| `CLIENT` | Outbound HTTP call wrapper (Feign, RestTemplate, WebClient) |
+| `CLIENT` | Outbound call wrapper (HTTP and messaging) |
 | `MAPPER` | Data mapper / converter |
 | `DTO` | Data transfer object — data carrier, no logic |
 | `OTHER` | Infrastructure / utility / framework / JDK / unclassified |

ast_java.py

@@ -83,7 +83,7 @@
 # Phase 11: `EDGE_SCHEMA` in `java_ontology.py` (canonical edge navigation schema; v14 re-index).
 # Phase 12: CALLS `callee_declaring_role`, supertype-walk dedup, pass3 unresolved counters (v15 re-index).
 # Bumps whenever extraction / enrichment semantics change.
-ONTOLOGY_VERSION = 15
+ONTOLOGY_VERSION = 16
 
 ROLE_ANNOTATIONS: dict[str, str] = {
     # Spring Web
@@ -2732,6 +2732,19 @@ def infer_role(annotation_names: Iterable[str]) -> str:
     return "OTHER"
 
 
+def _type_injects_messaging(type_decl: "TypeDecl") -> bool:
+    """True when the type injects a messaging template via field or constructor."""
+    for fld in type_decl.fields:
+        if fld.type_name in _INJECTED_TYPES_TO_CAPABILITY:
+            return True
+    for method in type_decl.methods:
+        if method.is_constructor:
+            for p in method.parameters:
+                if p.type_name in _INJECTED_TYPES_TO_CAPABILITY:
+                    return True
+    return False
+
+
 def infer_role_for_type(type_decl: "TypeDecl") -> str:
     """Role inference that also detects DTO-like passive data carriers.
 
@@ -2763,6 +2776,11 @@ def infer_role_for_type(type_decl: "TypeDecl") -> str:
         if name.endswith(suffix) and name != suffix:
             return "DTO"
 
+    # Types injecting messaging templates are outbound callers (CLIENT role),
+    # symmetric with CONTROLLER covering both HTTP and messaging inbound.
+    if _type_injects_messaging(type_decl):
+        return "CLIENT"
+
     return "OTHER"
 
 

docs/AGENT-GUIDE.md

@@ -14,7 +14,7 @@ Copy the block between `<!-- BEGIN` and `<!-- END` into your project's `AGENTS.m
 
 **Indexed content:** Java production sources plus SQL and YAML (use `search` `table`: `java`, `sql`, `yaml`, or `all`).
 
-**Ontology: 15** — if results look structurally wrong or empty across tools, the index may be missing, stale, or built with a different `ontology_version`; you cannot re-index via MCP — ask the operator to rebuild.
+**Ontology: 16** — if results look structurally wrong or empty across tools, the index may be missing, stale, or built with a different `ontology_version`; you cannot re-index via MCP — ask the operator to rebuild.
 
 **Responses:** On success, `search`, `find`, `describe`, `neighbors`, and `resolve` may include two top-level fields: `hints_structured` (≤5 suggested next-tool calls) and `advisories` (≤5 pure informational strings). Each `hints_structured` entry has `tool`, `args`, `actionable`, `label`, and `reason`. `actionable=true` means you can call the tool directly with `args`; `actionable=false` means partial/advisory — fill missing values or use as guidance. `reason` explains why the hint was emitted. `advisories` carry context education (fuzzy strategy warnings, role collision explanations, etc.) with no tool call suggestion. For `search`/`find`, echoed `limit`/`offset`. Hints are advisory; ignore them when `success` is false.
 
@@ -234,7 +234,7 @@ Returns **edges** with `attrs` (`confidence`, `strategy`, `match`, … on cross-
 | `COMPONENT` | General Spring component |
 | `CONFIG` | `@Configuration` class |
 | `ENTITY` | JPA / persistence entity |
-| `CLIENT` | Outbound HTTP call wrapper (Feign, RestTemplate, WebClient) |
+| `CLIENT` | Outbound call wrapper (HTTP and messaging) |
 | `MAPPER` | Data mapper / converter |
 | `DTO` | Data transfer object — data carrier, no logic |
 | `OTHER` | Infrastructure / utility / framework / JDK / unclassified |

skills/explore-codebase/SKILL.md

@@ -15,7 +15,7 @@ Any time you need to understand structure in an indexed Java codebase: locating
 
 **Indexed content:** Java production sources plus SQL and YAML (use `search` `table`: `java`, `sql`, `yaml`, or `all`).
 
-**Ontology: 15** — if results look structurally wrong or empty across tools, the index may be missing, stale, or built with a different `ontology_version`; you cannot re-index via MCP — ask the operator to rebuild.
+**Ontology: 16** — if results look structurally wrong or empty across tools, the index may be missing, stale, or built with a different `ontology_version`; you cannot re-index via MCP — ask the operator to rebuild.
 
 **Responses:** On success, `search`, `find`, `describe`, `neighbors`, and `resolve` may include two top-level fields: `hints_structured` (≤5 suggested next-tool calls) and `advisories` (≤5 pure informational strings). Each `hints_structured` entry has `tool`, `args`, `actionable`, `label`, and `reason`. `actionable=true` means you can call the tool directly with `args`; `actionable=false` means partial/advisory — fill missing values or use as guidance. `reason` explains why the hint was emitted. `advisories` carry context education (fuzzy strategy warnings, role collision explanations, etc.) with no tool call suggestion. For `search`/`find`, echoed `limit`/`offset`. Hints are advisory; ignore them when `success` is false.
 
@@ -235,7 +235,7 @@ Returns **edges** with `attrs` (`confidence`, `strategy`, `match`, … on cross-
 | `COMPONENT` | General Spring component |
 | `CONFIG` | `@Configuration` class |
 | `ENTITY` | JPA / persistence entity |
-| `CLIENT` | Outbound HTTP call wrapper (Feign, RestTemplate, WebClient) |
+| `CLIENT` | Outbound call wrapper (HTTP and messaging) |
 | `MAPPER` | Data mapper / converter |
 | `DTO` | Data transfer object — data carrier, no logic |
 | `OTHER` | Infrastructure / utility / framework / JDK / unclassified |

tests/test_client_role_rename.py

@@ -62,13 +62,15 @@ def test_no_legacy_feign_client_role_in_graph(tmp_path: Path) -> None:
     assert g.list_by_role("FEIGN_CLIENT") == []
 
 
-def test_resttemplate_class_unchanged(tmp_path: Path) -> None:
+def test_resttemplate_class_gets_client_role_from_messaging(tmp_path: Path) -> None:
     root = tmp_path / "proj"
     _copy_fixture(root)
     g = _build_graph(root, tmp_path / "g.kuzu")
     sym = _symbol_by_fqn(g.find_by_name_or_fqn("smoke.a.ClientA"), "smoke.a.ClientA")
     assert sym is not None
-    assert sym.role == "OTHER"
+    # ClientA injects KafkaTemplate → CLIENT role (symmetric with CONTROLLER)
+    assert sym.role == "CLIENT"
+    assert "MESSAGE_PRODUCER" in sym.capabilities
     assert "HTTP_CLIENT" not in sym.capabilities