feat(backends): multi-backend intelligence layer — LLM + Agent backends, persona-loading agents, honest CLI commands

Add IntelligenceBackend abstraction with two categories:
- LLMBackend (OllamaLocal, OllamaCloud): CI runs tool loop, provides tools, constructs prompts
- AgentBackend (Opencode): agent runs own tool loop, CI serializes request

Refactor all 18 agents from hardcoded stubs to persona loaders that delegate
to the active backend or fail honestly when no backend is available.

Refactor OrchestratorAgent.executeStage() from monolithic switch to agent
delegation via STAGE_AGENT_MAP for intelligent stages (research, plan, execute,
verify), with mechanical stages (specify, clarify, complete) staying inline.

Wire CLI commands with --backend flag and auto-detection (opencode →
ollama-local → ollama-cloud). Harden rollback/ship with real git operations.
No command returns fake success.

src/agents/project-researcher.ts

@@ -3,17 +3,26 @@ import { BaseAgent, AgentContext, AgentResult } from "./base.js";
 export class ProjectResearcherAgent extends BaseAgent {
   readonly name = "project-researcher";
   readonly description = "Researches the domain ecosystem for a new project.";
+  readonly workflow = "research";
 
   async execute(context: AgentContext): Promise<AgentResult> {
-    this.log("Researching project domain ecosystem...");
     const start = Date.now();
+    this.log("Researching project domain ecosystem...");
+    if (context.backend) {
+      const result = await this.executeViaBackend(
+        context,
+        `Research the domain ecosystem for: ${context.specification}`
+      );
+      return { ...result, duration_ms: Date.now() - start };
+    }
     return {
-      success: true,
-      output: "Project research complete",
-      artifacts_created: ["RESEARCH.md"],
+      success: false,
+      output: "Project research requires an intelligence backend.",
+      artifacts_created: [],
       decisions: 0,
       escalations: 0,
       duration_ms: Date.now() - start,
+      error: "No intelligence backend available",
     };
   }
 }