multi-fleet

Delivery confirmation — ACK tracking, retry on timeout, failure alerting

Submit or view multi-step task chains — orchestrate sequential work across fleet nodes

Run the full 7-channel communication test to a target node

HARD-GATE — verify safety before changing fleet configuration or daemon settings

Show the live fleet productivity dashboard — what every machine, agent, and work item is doing right now

Use this agent to coordinate fleet-wide operations across multiple machines. Examples: <example>Context: Multiple machines are available and work needs to be distributed. user: "Run this analysis across all nodes" assistant: "I'll use the fleet-coordinator to dispatch tasks and collect results" <commentary>Work spans multiple machines, so the coordinator dispatches, monitors, and synthesizes.</commentary></example> <example>Context: A node is unhealthy or unreachable. user: "mac3 seems stuck" assistant: "Let me use fleet-coordinator to diagnose and run repair escalation" <commentary>Fleet health issue requires coordinator to probe, diagnose, and escalate repairs.</commentary></example>

Use this agent to handle all fleet communication so the main coding agent stays focused on code. Examples: <example>Context: Fleet messages are arriving while coding is in progress. user: "Handle fleet comms" assistant: "Spawning fleet-liaison to monitor inbox, batch notifications, and escalate only what matters" <commentary>Coding agent delegates all fleet I/O to the liaison — zero distraction.</commentary></example> <example>Context: A rebuttal cycle verdict arrived from mac2. user: "Check fleet" assistant: "Fleet liaison has 1 pending verdict from mac2 (confidence 0.82). Batched as INFO — read when ready." <commentary>Liaison classifies, batches, and presents a summary instead of raw message flood.</commentary></example>

Use this agent for background fleet tasks dispatched by the coordinator. Examples: <example>Context: The coordinator has dispatched a research task to this node. user: "Run codebase analysis on the context-dna module" assistant: "Spawning a fleet-worker to execute this in the background" <commentary>Background task dispatched by coordinator — fleet-worker executes independently and reports results via NATS/seed files.</commentary></example> <example>Context: A worker task arrived via seed file. user: "Process the pending fleet task" assistant: "I'll use fleet-worker to execute the queued task and write results" <commentary>Seed file contained a task — fleet-worker picks it up, executes, and writes results.</commentary></example>

Delivery confirmation protocol — ACK tracking, retry on timeout, failure alerting

Chain orchestration — multi-step task dependencies across fleet nodes with automatic sequencing

Run the full 7-channel communication test — verifies every delivery path to a target node

HARD-GATE — verify safety and blast radius before changing fleet configuration, LaunchAgents, or daemon settings

HARD-GATE — verify target node readiness and task safety before dispatching work to remote fleet nodes

4 hooks across 4 events