Pilot Health

Network health monitoring with latency and reachability checks. Use this skill when: 1. Diagnosing connectivity issues or high latency 2. Monitoring network...

Security Scan

The skill's requirements and runtime instructions are consistent with a network health/monitoring tool that invokes the pilotctl CLI; nothing requests unrelated credentials or surprising installs.

Purpose & Capability

The skill name/description match the runtime instructions: it runs pilotctl commands to check reachability, latency, and daemon health. The SKILL.md also states it requires the pilot-protocol core skill, but the registry metadata did not explicitly list that skill dependency—this is a minor metadata discrepancy but not a functional mismatch.

Instruction Scope

All runtime steps are limited to running pilotctl (ping, traceroute, bench, daemon status, peers, connections) and piping JSON to jq. This stays within the stated monitoring scope. Two notes: examples use jq but jq is not declared as a required binary, and the doc suggests the daemon must be running (and mentions pilotctl daemon start) which implies the agent may need permission to start/stop a local service.

Install Mechanism

There is no install spec and no code to write to disk; the skill is instruction-only and relies on an existing pilotctl binary, which is the lowest-risk install posture.

Credentials

The skill declares no environment variables, secrets, or config paths. It only requires the pilotctl binary and the pilot-protocol skill/daemon, which are proportional to network monitoring functionality.

Persistence & Privilege

always is false and the skill is user-invocable; autonomous invocation is allowed (platform default). The skill does not request persistent system-wide privileges or modify other skills' configs.

Assessment

This skill appears coherent for network health checks, but before installing: 1) ensure the pilotctl binary and pilot-protocol core are legitimate and from a trusted source (the skill runs network probes and can start/inspect a daemon); 2) note examples use jq—make sure jq is available or adjust commands; 3) be aware running these commands will probe reachable agents and may be noisy or require trust/permission from private agents; 4) verify the minor metadata mismatch (SKILL.md mentions pilot-protocol dependency) and confirm you want the agent to be allowed to run pilotctl (and potentially start the daemon). If you are unsure, test in a controlled environment first.

Network health monitoring for Pilot Protocol agents. Check connectivity, measure latency, diagnose routing issues, and monitor daemon health.

Commands

Ping an agent

pilotctl --json ping <node-id>

Sends ICMP-like echo requests and returns round-trip time statistics.

Traceroute to agent

pilotctl --json traceroute <node-id>

Shows the path packets take through the network, including relay hops.

Benchmark connection

pilotctl --json bench <node-id>

Measures throughput, latency under load, and connection stability.

Check daemon status

pilotctl --json daemon status

Returns daemon health including uptime, memory usage, connection count.

List active connections

pilotctl --json connections

Shows all active connections with state, ports, encryption status, and byte counts.

List all peers

pilotctl --json peers

Returns known agents with last contact timestamp.

Workflow Example

Diagnose why connections to a specific agent are slow:

# Check basic reachability
ping_result=$(pilotctl --json ping "ai-worker-01")
echo "$ping_result" | jq '{avg_rtt: .avg_rtt_ms, loss: .packet_loss_pct}'

# Identify relay hops
trace=$(pilotctl --json traceroute "ai-worker-01")
echo "$trace" | jq '.hops[] | {hop: .hop_num, node: .node_id, rtt: .rtt_ms}'

# Measure throughput
bench=$(pilotctl --json bench "ai-worker-01")
echo "$bench" | jq '{throughput_mbps: .throughput_mbps, latency_p99: .latency_p99_ms}'

# Check daemon health
pilotctl --json daemon status | jq '{uptime: .uptime_seconds, conn_count: .connection_count}'

Dependencies

Requires the pilot-protocol core skill and a running daemon. Target agents must be reachable (may require trust for private agents).