Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

backend/apps/slack/events/message_posted.py (1)

83-92: Fix workspace resolution and add error handling for API call.

Two issues:

1. Line 83: event.get("team") may return None for IM conversations; Slack sometimes uses team_id, context_team_id, or user_team instead. The previous review flagged this as a critical blocker.
2. Line 84: client.conversations_info(channel=channel_id) can fail with API errors (rate limits, permissions, etc.), but it's not in a try-except, so failures crash the handler.

Apply the fix from the previous review to safely resolve workspace_id and wrap the API call:

-        workspace_id = event.get("team")
-        channel_info = client.conversations_info(channel=channel_id)
+        workspace_id = event.get("team") or event.get("team_id") or event.get("context_team_id") or event.get("user_team")
+        
+        if not workspace_id:
+            try:
+                channel_info = client.conversations_info(channel=channel_id)
+                channel_data = channel_info.get("channel", {})
+                workspace_id = (
+                    channel_data.get("team")
+                    or channel_data.get("context_team_id")
+                    or channel_data.get("user_team")
+                )
+            except Exception:
+                logger.exception("Failed to fetch workspace ID for DM.")
+                return
+        else:
+            try:
+                channel_info = client.conversations_info(channel=channel_id)
+            except Exception:
+                logger.exception("Failed to fetch channel info for DM.")
+                return
+        
+        if not workspace_id:
+            logger.error("Unable to determine workspace_id for DM channel %s.", channel_id)
+            return

backend/apps/slack/common/handlers/ai.py (1)

97-98: Handle None response before adding to context.

If rag_tool.query returns None, line 98 still calls add_to_context(query, None), which stores a user message without a bot response. While the method handles this case, it clutters the context with unanswered questions.

Only add to context if a response was generated:

     response = rag_tool.query(question=enhanced_query)
-    conversation.add_to_context(query, response)
+    if response:
+        conversation.add_to_context(query, response)
+    else:
+        logger.warning("RAG query returned no response for DM: %s", query[:50])
 
     return response

backend/apps/slack/events/message_posted.py (1)

92-92: Guard against missing id in channel_info["channel"]. Ensure channel_info.get("channel", {}).get("id") is present before calling Conversation.update_data to avoid a KeyError (e.g. bail or log if absent).
Location: backend/apps/slack/events/message_posted.py:92

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

backend/apps/slack/common/handlers/ai.py (1)

82-87: Log identifiers when DM context lookup fails.

We still emit a stack trace but lose the workspace_id/channel_id needed to debug these failures (and .exception is noisy for a handled condition). Switch to logger.error and include the identifiers so ops can trace the failing DM.

-    except (Workspace.DoesNotExist, Conversation.DoesNotExist):
-        logger.exception("Workspace or conversation not found for DM processing")
+    except (Workspace.DoesNotExist, Conversation.DoesNotExist):
+        logger.error(
+            "Workspace or conversation not found for DM processing: workspace_id=%s channel_id=%s",
+            workspace_id,
+            channel_id,
+        )

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro