iron_cage_types

Shared types and traits for the Iron Cage multi-agent safety framework.

Installation

[dependencies]
iron_cage_types = { version = "0.1", features = ["full"] }

Features

• enabled (default): Core functionality with all required dependencies
• full: All functionality (currently same as enabled)

Quick Start

use iron_cage_types::{Config, SafetyConfig, CostConfig};

// Create configuration for multi-agent system
let config = Config {
    safety: SafetyConfig {
        pii_detection_enabled: true,
        audit_log_path: Some("/var/log/safety.log".into()),
    },
    cost: CostConfig {
        budget_usd: 100.0,
        alert_threshold: 0.8,
    },
    reliability: Default::default(),
};

// Configuration is serializable when `enabled` feature is active
let json = serde_json::to_string(&config)?;

Documentation

• API Reference
• Architecture Guide

Responsibilities: Defines shared types, traits, and error types used across all iron_cage crates as foundation layer. Provides type-safe builders for configuration, comprehensive error handling with error_tools integration, and Serde serialization support. Requires Rust 1.75+, all platforms supported, serves as dependency for all other workspace crates.

In Scope:

• Common configuration types (SafetyConfig, CostConfig, ReliabilityConfig)
• Shared error types and Result aliases
• Core trait definitions (Agent, SafetyGuard, CostTracker, CircuitBreaker)
• Type-safe builders for configuration
• Serde serialization support (behind enabled feature)

Out of Scope:

• Implementation logic (see iron_cage_safety, iron_cage_cost, iron_cage_reliability)
• CLI interface (see iron_cage_cli)
• Python bindings (see iron_cage_cli)
• Business logic or orchestration

Source Files

Notes:

• Entries marked 'TBD' require manual documentation
• Entries marked '⚠️ ANTI-PATTERN' should be renamed to specific responsibilities

License

MIT