feat: real agent execution + real skill execution (system-prompt-driven)

- delegate_agent: now makes actual AI call with role-specific system prompts
  (coder=code review, architect=system design, devops=infrastructure)
- run_skill: now makes actual AI call with skill-specific system prompts
  (code_review, bug_fix, refactor, documentation, testing)
- Both return structured AI-generated results instead of placeholder text

src/bot/index.js

@@ -563,12 +563,50 @@ export async function initBot(config, api, tools, skills, agents) {
     delegate_agent: async (args) => {
       const agent = svc.agents.find(a => a.id === args.agent_id);
       if (!agent) return `❌ Agent not found: ${args.agent_id}`;
-      return `✅ Delegated to **${agent.name}**: "${args.task}"`;
+
+      // Actually execute the task with an agent-specific system prompt
+      const agentPrompts = {
+        coder: 'You are an expert code reviewer. Analyze the code for bugs, security issues, performance problems, and style improvements. Be specific and actionable.',
+        architect: 'You are a system architect. Analyze the request from an architecture perspective — patterns, scalability, maintainability, trade-offs. Provide clear recommendations.',
+        devops: 'You are a DevOps engineer. Focus on deployment, CI/CD, infrastructure, monitoring, and reliability. Provide practical, copy-paste-ready solutions.',
+      };
+
+      try {
+        const systemMsg = agentPrompts[agent.id] || `You are ${agent.name}. ${agent.description}`;
+        const agentMessages = [
+          { role: 'system', content: systemMsg },
+          { role: 'user', content: args.task },
+        ];
+        const result = await chatWithAI(agentMessages, { maxTokens: 4096 });
+        return `🤖 **${agent.name}**:\n${result}`;
+      } catch (e) {
+        return `❌ Agent ${agent.name} error: ${e.message}`;
+      }
     },
     run_skill: async (args) => {
       const skill = svc.skills.find(s => s.name === args.skill);
       if (!skill) return `❌ Skill not found: ${args.skill}`;
-      return `✅ Skill **${skill.name}** queued: ${args.input || '(none)'}`;
+
+      // Execute skill with a specialized system prompt
+      const skillPrompts = {
+        code_review: 'You are a code review expert. Analyze the provided code for: bugs, security vulnerabilities, performance issues, style problems, and improvements. Structure your review with severity levels (critical/warning/info).',
+        bug_fix: 'You are a debugging expert. Analyze the bug report and code, identify root cause, and provide a specific fix with explanation. Include edge cases.',
+        refactor: 'You are a refactoring specialist. Improve code quality while preserving behavior. Focus on: readability, DRY, SOLID principles, naming, and structure. Show before/after.',
+        documentation: 'You are a technical writer. Generate clear, comprehensive documentation from the provided code. Include: purpose, parameters, return values, examples, and edge cases.',
+        testing: 'You are a testing expert. Generate thorough tests for the provided code. Include: unit tests, edge cases, error cases. Use assertions with clear descriptions.',
+      };
+
+      try {
+        const systemMsg = skillPrompts[skill.name] || `You are a ${skill.name} expert. ${skill.description}`;
+        const skillMessages = [
+          { role: 'system', content: systemMsg },
+          { role: 'user', content: args.input || 'Please analyze the code and provide your expert review.' },
+        ];
+        const result = await chatWithAI(skillMessages, { maxTokens: 4096 });
+        return `📚 **${skill.name}**:\n${result}`;
+      } catch (e) {
+        return `❌ Skill ${skill.name} error: ${e.message}`;
+      }
     },
   };