//! Environment variable filtering for sandbox security.
//!
//! By default, the sandbox scrubs sensitive environment variables to prevent
//! credential leakage to untrusted MCP servers and agents. This module
//! implements the filtering logic described in ADR-001 and Phase 6 Hardening.
//!
//! # Security Model
//!
//! - **Scrub (Default)**: Remove known sensitive patterns (API keys, tokens, secrets)
//! - **Strict**: Only allow safe base variables (PATH, HOME, etc) + explicit allows
//! - **Passthrough**: Danger mode - pass all variables (not recommended)
//!
//! Additionally, "Execution Influence" variables (like LD_PRELOAD) can be stripped
//! independently or as part of Strict mode.
pub use ;
pub use matches_any_pattern;
pub use ;