Add agent explainability instrumentation and unify envelope field naming (#795)

Addresses recommendations from the UX developer's agent experience report.
Adds provenance predicates, DAG structure changes, error resilience, and
a published OWL ontology.

Explainability additions:

- Tool candidates: tg:toolCandidate on Analysis events lists the tools
  visible to the LLM for each iteration (names only, descriptions in config)
- Termination reason: tg:terminationReason on Conclusion/Synthesis events
  (final-answer, plan-complete, subagents-complete)
- Step counter: tg:stepNumber on iteration events
- Pattern decision: new tg:PatternDecision entity in the DAG between
  session and first iteration, carrying tg:pattern and tg:taskType
- Latency: tg:llmDurationMs on Analysis events, tg:toolDurationMs on
  Observation events
- Token counts on events: tg:inToken/tg:outToken/tg:llmModel on
  Grounding, Focus, Synthesis, and Analysis events
- Tool/parse errors: tg:toolError on Observation events with tg:Error
  mixin type. Parse failures return as error observations instead of
  crashing the agent, giving it a chance to retry.

Envelope unification:

- Rename chunk_type to message_type across AgentResponse schema,
  translator, SDK types, socket clients, CLI, and all tests.
  Agent and RAG services now both use message_type on the wire.

Ontology:

- specs/ontology/trustgraph.ttl — OWL vocabulary covering all 26 classes,
  7 object properties, and 36+ datatype properties including new predicates.

DAG structure tests:

- tests/unit/test_provenance/test_dag_structure.py verifies the
  wasDerivedFrom chain for GraphRAG, DocumentRAG, and all three agent
  patterns (react, plan, supervisor) including the pattern-decision link.

tests/unit/test_agent/test_agent_service_non_streaming.py

@@ -78,10 +78,10 @@ class TestAgentServiceNonStreaming:
 
         # Filter out explain events — those are always sent now
         content_responses = [
-            r for r in sent_responses if r.chunk_type != "explain"
+            r for r in sent_responses if r.message_type != "explain"
         ]
         explain_responses = [
-            r for r in sent_responses if r.chunk_type == "explain"
+            r for r in sent_responses if r.message_type == "explain"
         ]
 
         # Should have explain events for session, iteration, observation, and final
@@ -93,7 +93,7 @@ class TestAgentServiceNonStreaming:
         # Check thought message
         thought_response = content_responses[0]
         assert isinstance(thought_response, AgentResponse)
-        assert thought_response.chunk_type == "thought"
+        assert thought_response.message_type == "thought"
         assert thought_response.content == "I need to solve this."
         assert thought_response.end_of_message is True, "Thought message must have end_of_message=True"
         assert thought_response.end_of_dialog is False, "Thought message must have end_of_dialog=False"
@@ -101,7 +101,7 @@ class TestAgentServiceNonStreaming:
         # Check observation message
         observation_response = content_responses[1]
         assert isinstance(observation_response, AgentResponse)
-        assert observation_response.chunk_type == "observation"
+        assert observation_response.message_type == "observation"
         assert observation_response.content == "The answer is 4."
         assert observation_response.end_of_message is True, "Observation message must have end_of_message=True"
         assert observation_response.end_of_dialog is False, "Observation message must have end_of_dialog=False"
@@ -168,10 +168,10 @@ class TestAgentServiceNonStreaming:
 
         # Filter out explain events — those are always sent now
         content_responses = [
-            r for r in sent_responses if r.chunk_type != "explain"
+            r for r in sent_responses if r.message_type != "explain"
         ]
         explain_responses = [
-            r for r in sent_responses if r.chunk_type == "explain"
+            r for r in sent_responses if r.message_type == "explain"
         ]
 
         # Should have explain events for session and final
@@ -183,7 +183,7 @@ class TestAgentServiceNonStreaming:
         # Check final answer message
         answer_response = content_responses[0]
         assert isinstance(answer_response, AgentResponse)
-        assert answer_response.chunk_type == "answer"
+        assert answer_response.message_type == "answer"
         assert answer_response.content == "4"
         assert answer_response.end_of_message is True, "Final answer must have end_of_message=True"
         assert answer_response.end_of_dialog is True, "Final answer must have end_of_dialog=True"

tests/unit/test_agent/test_callback_message_id.py

@@ -29,7 +29,7 @@ class TestThinkCallbackMessageId:
 
         assert len(responses) == 1
         assert responses[0].message_id == msg_id
-        assert responses[0].chunk_type == "thought"
+        assert responses[0].message_type == "thought"
 
     @pytest.mark.asyncio
     async def test_non_streaming_think_has_message_id(self, pattern):
@@ -58,7 +58,7 @@ class TestObserveCallbackMessageId:
         await observe("result", is_final=True)
 
         assert responses[0].message_id == msg_id
-        assert responses[0].chunk_type == "observation"
+        assert responses[0].message_type == "observation"
 
 
 class TestAnswerCallbackMessageId:
@@ -74,7 +74,7 @@ class TestAnswerCallbackMessageId:
         await answer("the answer")
 
         assert responses[0].message_id == msg_id
-        assert responses[0].chunk_type == "answer"
+        assert responses[0].message_type == "answer"
 
     @pytest.mark.asyncio
     async def test_no_message_id_default(self, pattern):

tests/unit/test_agent/test_orchestrator_provenance_integration.py

@@ -69,7 +69,7 @@ def collect_explain_events(respond_mock):
     events = []
     for call in respond_mock.call_args_list:
         resp = call[0][0]
-        if isinstance(resp, AgentResponse) and resp.chunk_type == "explain":
+        if isinstance(resp, AgentResponse) and resp.message_type == "explain":
             events.append({
                 "explain_id": resp.explain_id,
                 "explain_graph": resp.explain_graph,

tests/unit/test_agent/test_parse_chunk_message_id.py

@@ -20,7 +20,7 @@ class TestParseChunkMessageId:
 
     def test_thought_message_id(self, client):
         resp = {
-            "chunk_type": "thought",
+            "message_type": "thought",
             "content": "thinking...",
             "end_of_message": False,
             "message_id": "urn:trustgraph:agent:sess/i1/thought",
@@ -31,7 +31,7 @@ class TestParseChunkMessageId:
 
     def test_observation_message_id(self, client):
         resp = {
-            "chunk_type": "observation",
+            "message_type": "observation",
             "content": "result",
             "end_of_message": True,
             "message_id": "urn:trustgraph:agent:sess/i1/observation",
@@ -42,7 +42,7 @@ class TestParseChunkMessageId:
 
     def test_answer_message_id(self, client):
         resp = {
-            "chunk_type": "answer",
+            "message_type": "answer",
             "content": "the answer",
             "end_of_message": False,
             "end_of_dialog": False,
@@ -54,7 +54,7 @@ class TestParseChunkMessageId:
 
     def test_thought_missing_message_id(self, client):
         resp = {
-            "chunk_type": "thought",
+            "message_type": "thought",
             "content": "thinking...",
             "end_of_message": False,
         }
@@ -64,7 +64,7 @@ class TestParseChunkMessageId:
 
     def test_answer_missing_message_id(self, client):
         resp = {
-            "chunk_type": "answer",
+            "message_type": "answer",
             "content": "answer",
             "end_of_message": True,
             "end_of_dialog": True,