Monitoring what your agents are doing just got significantly more powerful. This release adds a dedicated execution view for complex plans, richer activity feeds, and real performance data — so you can understand not just what happened, but how well it went.
Execution Layout Modal For Processes And Projects
Clicking the main entity card on a Process or Project detail view now opens a full execution layout modal. It presents a side-by-side view: the execution tree on the left and run history on the right.
- Flat grid layout — Every entity in the hierarchy (root, tasks, subtasks) occupies its own aligned row, making the full execution scope visible at a glance
- Run carousel — Navigate between multiple runs for each entity using color-coded status chips (green for success, red for failure, blue for running, yellow for pending)
- Collapsible subtasks — Expand or collapse subtask groups to focus on the level of detail you need
- Execution mode badges — See at a glance whether each entity runs via BASH, PYTHON, LLM, ORCHESTRATE, or an interactive mode
- Open from activity — You can also launch the execution modal directly from activity run entries
Attachment Viewer
If a run has associated files (logs, reports, images), an attachment indicator appears in the execution tree. Clicking it opens a viewer that renders markdown and text inline and provides downloads for other file types.
Activity Feed: Flat And Hierarchical Views
The activity panel now offers two ways to look at run history:
- Flat view — All runs in chronological order, regardless of which entity they belong to. Useful for "what happened recently?"
- Hierarchical view — Runs grouped under their parent entities, preserving the planning structure. Useful for "how is this plan progressing?"
A toggle in the panel header lets you switch between views. Run status chips provide a compact visual timeline underneath each run display, ordered chronologically from left to right.
Real Performance Metrics
Planning entity detail views now show actual performance data instead of placeholders. Metrics are loaded on demand — they appear when you open the detail card and update without slowing down navigation.
This includes timing data, success rates, and run counts calculated from real execution history, giving you a factual picture of how each part of your plan has been performing.