Check the health of the axon agent runtime itself: whether the collector is running, DB is accessible, and alert dispatch is configured. Returns health status and diagnostic messages.

Get battery state (percentage, is_charging, estimated_minutes_remaining). Returns null/false on non-battery platforms (returns {ok: false} gracefully).

Capture GPU state (memory usage, temperature, utilization). Returns {ok: false} gracefully on systems without GPU or on unsupported platforms.

Query historical snapshots over a time range (last_1h, last_6h, last_24h) at specified intervals (5m, 15m, 1h). Returns time series of CPU %, memory, temperature, and anomaly scores.

Capture current hardware state: CPU %, memory usage, die temperature, throttling status, and anomaly score. Returns JSON with cpu_usage_pct, ram_used_gb, ram_total_gb, die_temp_celsius, throttling, and anomaly_score.

Identify the process(es) responsible for CPU and memory anomalies. Returns culprit process with PID, command, CPU %, RAM GB, anomaly score, impact level, and recommended fix.

Report overall session health: impact level (healthy, degrading, strained, critical), RAM pressure tier, thermal state, and count of stale axon instances.

Profile the system once (CPU count, RAM total, product name, macOS version). Used during setup/initialization.

Request advice on safe parallelism and risk level given a workload type and requested parallelism. Returns safe_parallelism (capped based on current load), risk_narrative, and estimated_minutes_to_recovery.