flywheel-swarm-status

Inspect swarm agent status, bead progress, and inbox messages to detect stalled work.

3|1|Updated Apr 15, 2026
One-click install
npx skills add https://github.com/burningportra/agent-flywheel-plugin --skill flywheel-swarm-status-burningportra
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: flywheel-swarm-status
Source: https://github.com/burningportra/agent-flywheel-plugin/tree/main/skills/flywheel-swarm-status
Command: npx skills add https://github.com/burningportra/agent-flywheel-plugin --skill flywheel-swarm-status-burningportra

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When running multi-agent coding swarms, it is hard to know which agents are still working, which beads are stuck, and whether the swarm has finished. This Skill consolidates bead status, agent inbox messages, and todo state into one status report with a recommended next action. ## Core Features & Use Cases - Bead Status Table: Runs br list --json and renders an ID/Title/Status/Updated table, highlighting beads that are in_progress. - Agent Inbox Check: Calls fetch_inbox via the agent-mail MCP tool to surface messages from running agents. - Stuck Detection: Flags beads whose updated_at is older than 30 minutes while still in_progress, and recommends stopping and restarting the swarm. - Completion Routing: When all swarm beads finish, routes into the wave-completion review flow (risky-bead detection, review options, flywheel_review) instead of ending prematurely. - Use Case: Midway through a parallel implementation wave, ask for a status check to see which agents are stalled, read their latest messages, and decide whether to restart the swarm or proceed to review. ## Quick Start Ask the agent to check the current swarm status and tell you which beads are stuck or finished.

Frequently Asked Questions about flywheel-swarm-status

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I check the status of running swarm agents?

Run the swarm status check, which executes br list --json to build a bead status table and calls fetch_inbox through the agent-mail MCP tool to show messages from running agents. Beads still in_progress are highlighted.

How to detect stuck or stalled beads in a swarm?

A bead is flagged as stuck when its updated_at timestamp is older than 30 minutes while its status remains in_progress. The status report highlights these beads and recommends stopping and restarting the swarm.

What happens when all swarm beads have finished?

The workflow does not end at swarm completion. It reads the review instructions and runs the wave-completion review: risky-bead detection, a consolidated review choice prompt, and routing through flywheel_review.

Can I use this status check without the agent-mail MCP tool?

The inbox portion depends on the agent-mail MCP tool's fetch_inbox function to display agent messages. Without it, only the br list bead table and todo status portions of the report are available.

Why does the status check recommend restarting the swarm?

It recommends a restart only when stuck agents are detected, meaning beads stayed in_progress without updates for over 30 minutes. Restarting clears stalled agents so work can resume cleanly.