Participants: Marcos Pagnucco (Kindo), Nathan Ellis (Deloitte), Sumanth Tadikonda, Harish, Harshal Vasudeo, and ~40 others
🔥 ThreatConnect MCP - ~80% Failure Rate - Every time a tool call is made to ThreatConnect, ~80% of calls fail because parameters are rejected. System retries/brute-forces until one succeeds. Something wrong with MCP parameter configuration or description. Escalated to engineering.
🔥 Integration Routing - Cannot Select Specific Integration - When multiple integrations of the same type are configured (e.g., 3 Jiras, 3 CrowdStrikes), there is no way to specify which one an agent should use. Long-standing request tracked by Mo - no status update provided for weeks. Escalated with priority.
🔥 PM Visibility Gap - No Updates for 2 Weeks - Primary PM (Mo) absent for 2 weeks. No status updates provided to Deloitte. Centralized priority tracking dashboard promised over a month ago was never shown. Multiple teams with different priorities and no visibility into what's being tracked. Nathan: "I'm getting a little frustrated because we've been waiting for updates."
🔧 Sandbox Race Condition - Fix in Testing - Root cause found: race condition in cleanup process causing EBS unmount failures. Deep debugging session with Nathan captured critical logs. Fix in testing - expected tomorrow morning at latest. Instance already upgraded to v3.16.
⚠️ Agent UI - Completion Still Not Reflected - Agents complete on backend but UI doesn't update. Users must refresh manually 2-3 times. Front-end polling/streaming issue persists despite backend improvements.
⚠️ Jira DC Triggers Non-Functional - Custom Jira Data Center triggers don't initiate agents at all on ITS. Multiple filter configurations attempted. Lower priority than integration routing per Deloitte.
📦 SailPoint ISC Write Operations - Requested 1-1.5 months ago. Jira created by Mo, no status update provided. Currently read-only - write operations needed.
📦 Canvas Disclaimer Customization - Separate disclaimers for Canvas and chat. Implemented, in next release (same release as sandbox fix).
📦 Agent Failure Notes - Better visibility into why agents failed. Engineering has discussed; expected in next release.
💡 Agent Output UI Enhancements (New request - Low Priority) - Tool call output areas are cluttered and not readable. Request for resizable windows for prompt/output areas. Also: show which asset and user ID was used during tool calls.
📦 SMK Deployment Automation - Preflight script with new pre-flight functionality ready. Marcos to send package + instructions to Deloitte within hours. Validates cluster networking, connection strings, and access before deployment.
📅 Meeting Cadence - Out-of-band meeting proposed for next week to bring full updates that Marcos couldn't provide. Mo expected to provide personal update to Deloitte.