feat: database layer + canvas validator/converter + LLM inference activities

- Add pkg/db models and CRUD methods for workflows
- Add internal/routing canvas validator (DAG check, connectivity)
- Add internal/routing canvas converter (Canvas → WorkflowSpec)
- Register LLMInferenceActivity and LLMBatchInferenceActivity
- Update api/server and cmd/server with database integration
- Add K8s environment variable support
- Update activity knowledge base with LLM activities
- Add .env.example configuration template

internal/routing/activity_knowledge_base.json

@@ -473,10 +473,130 @@
         "dependencies": [],
         "notes": "Must run before LLM Router to provide auth token. Call early in workflow."
       }
+    },
+    {
+      "name": "LLMInferenceActivity",
+      "description": "Call LLM API with custom prompt and get response text",
+      "category": "llm",
+      "inputs": {
+        "model": {
+          "type": "string",
+          "description": "Model ID (reasoning, ornith:35b, ornith:13b, qwen2.5:3b)",
+          "required": true,
+          "examples": ["reasoning", "ornith:35b"]
+        },
+        "system_prompt": {
+          "type": "string",
+          "description": "System instruction for the model",
+          "required": false,
+          "default": ""
+        },
+        "user_prompt": {
+          "type": "string",
+          "description": "User message to send to the model",
+          "required": true
+        },
+        "temperature": {
+          "type": "number",
+          "description": "Sampling temperature (0.0-1.0, higher=more creative)",
+          "required": false,
+          "default": 0.7
+        },
+        "max_tokens": {
+          "type": "integer",
+          "description": "Maximum tokens in response",
+          "required": false
+        }
+      },
+      "outputs": {
+        "response": {
+          "type": "string",
+          "description": "LLM response text"
+        },
+        "model": {
+          "type": "string",
+          "description": "Model used for inference"
+        },
+        "stop_reason": {
+          "type": "string",
+          "description": "Why inference stopped (stop_sequence, length, etc)"
+        },
+        "tokens_used": {
+          "type": "integer",
+          "description": "Total tokens consumed"
+        }
+      },
+      "constraints": {
+        "defaultTimeout": "120s",
+        "isFlaky": true,
+        "recommendedRetries": 2,
+        "retryBackoff": 2.0,
+        "dependencies": [],
+        "notes": "API-dependent. Network flaky. Use for single prompts. See LLMBatchInferenceActivity for multiple."
+      }
+    },
+    {
+      "name": "LLMBatchInferenceActivity",
+      "description": "Call LLM API multiple times sequentially with different prompts",
+      "category": "llm",
+      "inputs": {
+        "model": {
+          "type": "string",
+          "description": "Model ID (reasoning, ornith:35b, ornith:13b, qwen2.5:3b)",
+          "required": true
+        },
+        "system_prompt": {
+          "type": "string",
+          "description": "System instruction (same for all prompts)",
+          "required": false
+        },
+        "prompts": {
+          "type": "array",
+          "description": "List of user prompts to process",
+          "required": true,
+          "items": {
+            "type": "string"
+          }
+        },
+        "temperature": {
+          "type": "number",
+          "description": "Sampling temperature (0.0-1.0)",
+          "required": false,
+          "default": 0.7
+        }
+      },
+      "outputs": {
+        "responses": {
+          "type": "array",
+          "description": "List of LLM responses (parallel to input prompts)",
+          "items": {
+            "type": "string"
+          }
+        },
+        "model": {
+          "type": "string",
+          "description": "Model used"
+        },
+        "errors": {
+          "type": "array",
+          "description": "Error messages for failed prompts",
+          "items": {
+            "type": "string"
+          }
+        }
+      },
+      "constraints": {
+        "defaultTimeout": "600s",
+        "isFlaky": true,
+        "recommendedRetries": 1,
+        "retryBackoff": 2.0,
+        "dependencies": [],
+        "notes": "Sequential processing of multiple prompts. Use for batch analysis, summarization, etc."
+      }
     }
   ],
   "metadata": {
-    "totalActivities": 10,
+    "totalActivities": 12,
     "lastUpdated": "2025-08-31T00:00:00Z",
     "categories": {
       "repository": 1,
@@ -488,7 +608,8 @@
       "approval": 1,
       "storage": 1,
       "memory": 1,
-      "authentication": 1
+      "authentication": 1,
+      "llm": 2
     }
   }
 }