Agent Platform: add conv_search to MCP tool examples

Trio now shows three distinct retrieval scopes:
- get_context — assembled context window for one participant
- knowledge_search — semantic memory / knowledge graph
- conv_search — search actual conversations across the entire store

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: rmdevpro

open-source/agent-platform.html

@@ -109,7 +109,7 @@ <h3>Modular Hot Loading</h3>
 
       <h3>Interfaces</h3>
       <ul>
-        <li><b>MCP tool interface</b> — ten-plus tools agents can call directly, such as <code>get_context</code> (assemble and return an agent's context window) and <code>knowledge_search</code> (search semantic memory and the knowledge graph)</li>
+        <li><b>MCP tool interface</b> — ten-plus tools agents can call directly, such as <code>get_context</code> (assemble and return an agent's context window), <code>knowledge_search</code> (search semantic memory and the knowledge graph), and <code>conv_search</code> (search actual conversations across the entire store)</li>
         <li><b>OpenAI-compatible chat endpoint</b> — the primary access path to the Imperator, provided via the industry-standard OpenAI-compatible endpoint protocol</li>
         <li><b>Bundled chatbot UI</b> — a lightweight web interface shipped with the Broker for talking to the Imperator directly</li>
       </ul>