Module security 

Security subsystem for policy enforcement, sandboxing, and secret management.

Re-exports

pub use audit::AuditEvent;

pub use audit::AuditEventType;

pub use audit::AuditLogger;

pub use detect::create_sandbox;

pub use detect::linux_memcg_available;

pub use detect::SandboxPosture;

pub use detect::sandbox_posture;

pub use estop::EstopLevel;

pub use estop::EstopManager;

pub use estop::EstopState;

pub use estop::ResumeSelector;

pub use external_content::ContentSafety;

pub use external_content::FramingPolicy;

pub use external_content::OutboundPolicy;

pub use external_content::ScanOutcome;

pub use external_content::ScanPolicy;

pub use external_content::ScreenVerdict;

pub use external_content::cap_untrusted;

pub use external_content::frame_untrusted;

pub use external_content::new_marker_id;

pub use external_content::sanitize_untrusted;

pub use external_content::scan_untrusted;

pub use external_content::scrub_outbound;

pub use ingress::IngressPolicy;

pub use ingress::ingress_policy;

pub use otp::OtpValidator;

pub use traits::NoopSandbox;

pub use traits::Sandbox;

pub use iam_policy::IamPolicy;

pub use iam_policy::PolicyDecision;

pub use nevis::NevisAuthProvider;

pub use nevis::NevisIdentity;

pub use leak_detector::LeakDetector;

pub use leak_detector::LeakResult;

pub use prompt_guard::GuardAction;

pub use prompt_guard::GuardResult;

pub use prompt_guard::PromptGuard;

Modules

audit

Audit logging for security events Each audit entry is chained via a Merkle hash: entry_hash = SHA-256(prev_hash || canonical_json). This makes the trail tamper-evident — modifying any entry invalidates all subsequent hashes.

auth_provider

RFC inbound authentication seam: the AuthProvider trait + a default-deny ProviderRegistry.

detect

Auto-detection of available security features

docker

Docker sandbox (container isolation)

domain_matcher

estop

external_content

Content-safety helpers for untrusted external SOP payloads.

firejail

Firejail sandbox (Linux user-space sandboxing) Firejail is a SUID sandbox program that Linux applications use to sandbox themselves.

iam_policy

IAM-aware policy enforcement for Nevis role-to-permission mapping. Evaluates tool and workspace access based on Nevis roles using a deny-by-default policy model. All policy decisions are audit-logged.

ingress

Universal ingress policy front door (RFC phase 1).

leak_detector

Credential leak detection for outbound content.

nevis

Nevis IAM authentication model_provider for ZeroClaw.

otp

pairing

playbook

Incident response playbook definitions and execution engine.

policy

prompt_guard

Prompt injection defense layer.

secrets

traits

Sandbox trait for pluggable OS-level isolation.

vulnerability

Vulnerability scan result parsing and management. Parses vulnerability scan outputs from common scanners (Nessus, Qualys, generic CVSS JSON) and provides priority scoring with business context adjustments.

Structs

DomainMatcher

PairingGuard

SecretStore

Manages encrypted storage of secrets (API keys, tokens, etc.)

SecurityPolicy

Enums

AutonomyLevel

The agent’s autonomy level, ordered from least to most autonomous.

Functions

redact

Redact sensitive values for safe logging. Shows first 4 characters + “***” suffix. Uses char-boundary-safe indexing to avoid panics on multi-byte UTF-8 strings. This function intentionally breaks the data-flow taint chain for static analysis.

scrub

Scrub credential leaks from arbitrary text before it crosses into a log record or any other sink. Routes through the global LeakDetector so every known credential shape is redacted in one place rather than via per-callsite regexes. Clean input is returned unchanged.

scrub_with_config

Scrub credential leaks using the configured leak-detection policy.