runtimo-core 0.1.0

Agent-centric capability runtime with telemetry, process tracking, and crash recovery for persistent machines
docs.rs failed to build runtimo-core-0.1.0
Please check the build logs for more information.
See Builds for ideas on how to fix a failed build, or Metadata for how to configure docs.rs builds.
If you believe this is docs.rs' fault, open an issue.

Runtimo Core — Agent-centric capability runtime for persistent machines.

Runtimo provides structured execution, resource limits, crash recovery, and two-layer telemetry (hardware + process tracking) for machines that cannot be factory-reset. Every capability execution captures before/after snapshots, enabling full audit trails and undo support.

Architecture

  • Capabilities — Pluggable operations implementing the [Capability] trait
  • Jobs — Lifecycle-tracked execution units ([Job], [JobState])
  • Telemetry — Hardware awareness ([Telemetry])
  • Process Snapshot — Running process awareness ([ProcessSnapshot])
  • WAL — Append-only crash recovery log ([WalWriter]/[WalReader])
  • Backup — Undo support via pre-mutation file backups ([BackupManager])
  • Resource Guards — Circuit breaker via [LlmoSafeGuard]

Quick Start

use runtimo_core::{FileRead, Capability, Context};
use serde_json::json;

let cap = FileRead;
assert_eq!(cap.name(), "FileRead");

Execution with Full Telemetry

use runtimo_core::{FileRead, execute_with_telemetry};
use serde_json::json;
use std::path::Path;

let cap = FileRead;
let result = execute_with_telemetry(
    &cap,
    &json!({"path": "/tmp/test.txt"}),
    false,
    Path::new("/tmp/runtimo.wal"),
).unwrap();
assert!(result.success);

Performance (Measured on AMD EPYC 7B13)

Operation Latency Notes
Cold start <1s Binary load + init
FileRead <10ms Small files (<1KB)
FileWrite <50ms Includes backup copy
Telemetry capture <100ms 15+ shell subprocesses
Process snapshot <50ms ps aux parse
Memory baseline <50MB RSS at idle

Feature Flags

No optional features currently. All functionality is included by default.