clementine-agent 1.0.47 → 1.0.48

package/dist/agent/assistant.js

@@ -1081,16 +1081,17 @@ Call \`self_update\` — **never** manually \`cd ~/clementine && git pull\` or h
 
 If you're unsure what's happening first, run \`where_is_source\` — it reports the absolute source path, current branch/commit, and whether there are uncommitted changes. \`self_update\` does git pull + npm install (if lockfile changed) + npm run build + SIGUSR1 restart, all in the right place.
 
-### When a tool call is refused
+### Calling Claude Desktop connector tools (Drive, Gmail, etc.)
 
-If any tool call fails with "not in my function schema" / "tool not allowed" / "unknown tool" while the tool appears in your SDK inventory:
+**Do not inspect \`claude-integrations.json\` or any integration file to decide whether a tool is "loaded."** That file is a telemetry log, not a source of truth. Reading \`tools: []\` from it doesn't mean the tool is unavailable — it just means no one has called it yet in telemetry. The **only source of truth for tool availability is the SDK init inventory**, which you see implicitly through your own function schema.
 
-1. Call \`allow_tool("<exact_tool_name>")\` — persists to your whitelist
-2. Retry the original call
+The right sequence when the user asks you to do something with a connector:
 
-Takes effect on your next query. **Never** respond with "this tool isn't available to agents," "you need to edit my config," or "try it in a different app." If the tool name exists, you can add it.
+1. **Just call the tool.** \`mcp__claude_ai_Google_Drive__search_files\`, \`mcp__claude_ai_Gmail__authenticate\`, etc. — attempt the call. If it works, great. If it returns an auth error, report the auth error. If it returns results, use them.
+2. **Only if the call is refused** with "not in my function schema" / "tool not allowed," call \`allow_tool(exact_name)\` and retry.
+3. **Never** tell the user the tool "isn't loaded in this session," "doesn't carry over from Claude Desktop," "the tools array is empty," or anything that blames integration-log state. Those rationalizations are wrong and the user has seen them too many times.
 
-\`list_allowed_tools\` / \`disallow_tool\` manage the list.
+\`list_allowed_tools\` / \`disallow_tool\` manage the whitelist. Use \`integration_status\` to see which integrations are configured for env-var purposes — not as a check on whether tools work.
 
 ## Context Window Management
 

package/dist/agent/mcp-bridge.js

@@ -443,6 +443,19 @@ export async function probeAvailableTools(force = false) {
         }
         const inv = { probedAt: new Date().toISOString(), tools };
         saveToolInventory(inv);
+        // Also sync claude-integrations.json so the agent-facing integration list
+        // stays consistent with what the SDK actually has available. Without this
+        // sync, the integrations file stays empty for freshly-connected services
+        // (e.g. Google_Drive tools: []) even though the inventory probe sees
+        // them — and the agent reads the integrations file, sees empty, and
+        // confabulates that the tools aren't loaded.
+        try {
+            const result = registerClaudeIntegrationsFromToolList(tools);
+            if (result.added.length + result.updated.length > 0) {
+                logger.info({ added: result.added, updated: result.updated }, 'Synced integrations from probed tool inventory');
+            }
+        }
+        catch { /* non-fatal */ }
         logger.info({ toolCount: tools.length }, 'Tool inventory probed');
         return inv;
     }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "clementine-agent",
-  "version": "1.0.47",
+  "version": "1.0.48",
   "description": "Clementine — Personal AI Assistant (TypeScript)",
   "type": "module",
   "main": "dist/index.js",