Changelog — Week of March 30th, 2026
New features and improvements
Agents & Tools
Added agent time logging — agents now track estimated time spent resolving each ticket, feeding into a new performance overview that shows how much manual work the agent is handling.
Added search query tool — agents can now run direct search and aggregation queries across all synced tickets and documentation for richer analysis and reporting.
Agents running custom scripts can now make multiple API calls in parallel (PSA, Microsoft Graph, documentation providers), significantly speeding up bulk operations like multi-tenant audits.
Agent operations that list end-client companies no longer cap at 15 results — all companies are now accessible via pagination.
Workflows & Triggers
Added per-workflow credit usage — the workflows list now shows how many credits each workflow has used this month and how many unique tickets it handled.
Triage workflows can now assign closed or cancelled statuses to tickets, useful for spam filtering or auto-closing duplicates.
When one agent triggers another workflow, the target now runs immediately regardless of its configured schedule window — previously, chained workflows could be skipped if triggered outside business hours.
Autotask
Technician expertise profiles now include tickets where the tech was a secondary resource (collaborator), improving dispatch accuracy for Autotask users.
Bug fixes
Sandbox
Fixed an issue where sandbox bash commands failed with a timeout error after 60 seconds regardless of the configured timeout limit.
Agents
Fixed agents re-processing their own ticket changes and missing configuration/asset updates between runs due to incomplete ticket hashing.
ConnectWise
Fixed ticket triage failing entirely on Catch-All company tickets due to a rejected flag toggle.
Hudu
Fixed stale Hudu documents persisting in search results when a synced category returned zero documents.
Autotask
Fixed logging time to Autotask tickets failing when the agent didn't include a customer-facing message.
RMM
Fixed RMM scripts failing with a 404 error when large batches queued up, causing the script to expire before the RMM agent could fetch it.