roboticus-core 0.9.8

Shared types, config parsing, personality system, and error types for the Roboticus agent runtime
Documentation

roboticus-core

Version 0.5.0

Shared types, configuration parsing, encrypted credential storage, personality system, and error types for the Roboticus agent runtime.

This is the leaf crate in the dependency graph -- every other Roboticus crate depends on it, and it depends on no internal crates.

Key Types & Traits

Type Module Description
RoboticusConfig config Top-level configuration parsed from roboticus.toml
RoboticusError / Result error Unified error type (13 variants) used across all crates
Keystore keystore Encrypted key-value store for API keys and secrets
SurvivalTier types Financial health tier (T1--T4) derived from balance
AgentState types Agent lifecycle state
ApiFormat types LLM provider API format (OpenAI, Ollama, Google, Anthropic)
ModelTier types Model capability tier
PolicyDecision types Allow / Deny / Escalate
RiskLevel types Tool risk classification
SkillManifest types Skill metadata and trigger configuration
Theme style Terminal theme (CRT, orange, green)

Usage

[dependencies]
roboticus-core = "0.5"

use roboticus_core::{RoboticusConfig, RoboticusError, Result, Keystore};

// Load configuration
let config = RoboticusConfig::from_file("roboticus.toml")?;

// Access encrypted credentials
let ks = Keystore::new(Keystore::default_path());
if let Some(key) = ks.get("openai_api_key") {
    println!("Key loaded from keystore");
}

Documentation

API docs are available on docs.rs.

License

Licensed under Apache-2.0. See LICENSE for details.