Visualize distributed agent workflow execution state and liveness from run records.
npx skills add https://github.com/shawnroos/auto --skill auto-watch
Please help me install this Agent Skill. Skill: auto-watch Source: https://github.com/shawnroos/auto/tree/main/skills/auto-watch Command: npx skills add https://github.com/shawnroos/auto --skill auto-watch
This skill solves the lack of visibility into complex, multi-step agent workflows by providing a real-time, hierarchical view of dispatched tasks, helping operators identify wedged or stalled processes.
Use the auto-watch skill to render the current agent tree for the active run by providing the run identifier to the monitoring script.
High-intent search queries and answers about installing and using this skill.
To monitor distributed agent workflows, this skill maps run-record structures against active task process data to visualize execution state and identify stalled or unresponsive nodes in real-time. It provides a hierarchical view of dispatched tasks to detect wedged processes.
Visualizing live agent execution trees is done by rendering a clear hierarchical map of the driver, work steps, and nested fan-out agents. This allows operators to supervise complex, multi-step agent workflows by overlaying real-time process status from the harness.
Agent liveness monitoring detects dead or hung agents by overlaying real-time process status from the harness onto the execution tree. It automatically flags nodes that have exceeded their defined age threshold, enabling proactive supervision of silently unresponsive tasks.
Yes, you need access to the run-record and system task monitoring tools to supervise agent processes. The skill requires these dependencies to reconcile process reality with the agent's internal state and accurately identify stalled nodes.
To render the current agent tree for an active run, provide the run identifier to the monitoring script. This initiates the visualization of the execution state and liveness of your distributed agent workflow.
khgs2411Read DASHBOARD.md to report Flow project status and next steps.
openeuler-mirrorCorrelate kernel wait channels with user-space thread states to diagnose hung Linux processes.
zylos-aiMonitor Claude agent status and restart it after crashes using PM2.
imMamdouhaboammarDesign multi-agent AI pipelines with distributed systems engineering principles.
