Crate aa_core 

Core domain logic for Agent Assembly.

This crate is no_std compatible and contains the foundational types, traits, and pure logic shared across all other crates in the workspace. It has no runtime or I/O dependencies.

Feature Flags

• std (default): enables std-dependent convenience impls (e.g. From<SystemTime>)
• alloc: enables heap types (String, Vec, BTreeMap) in no_std environments
• serde: enables Serialize/Deserialize derives on all core types (added in AAASM-22–25)
• test-utils: exposes PermitAllEvaluator and DenyAllEvaluator for downstream test code
• std (also default): enables CredentialScanner and all std-dependent types
• alloc (also default via std): enables AuditEntry, AuditEventType, and all audit types

Re-exports

pub use dev_tool::GovernanceLevel;

pub use identity::AgentId;

pub use identity::SessionId;

pub use policy::EnforcementMode;

pub use policy::FileMode;

pub use policy::PolicyDecision;

pub use policy::PolicyError;

pub use risk_tier::RiskTier;

pub use agent::AgentContext;

pub use agent::AgentContextBuilder;

pub use approval::ApprovalKind;

pub use dev_tool::DevToolKind;

pub use dev_tool::McpServerInfo;

pub use dev_tool::AdapterError;

pub use dev_tool::DevToolAdapter;

pub use dev_tool::DevToolInfo;

pub use llm::Model;

pub use llm::Provider;

pub use policy::ArgsJson;

pub use policy::GovernanceAction;

pub use policy::PolicyDocument;

pub use policy::PolicyEvaluator;

pub use policy::PolicyResult;

pub use policy::PolicyRule;

pub use audit::AuditEntry;

pub use audit::AuditEventType;

pub use audit::AuditLog;

pub use audit::AuditLogError;

pub use audit::Lineage;

pub use capability::action_to_capability;

pub use capability::capability_is_denied;

pub use capability::merge_capabilities;

pub use capability::Capability;

pub use capability::CapabilitySet;

pub use capability::EffectivePermissions;

pub use capability::PermissionSource;

pub use config::AgentConnectConfig;

pub use config::ConfigError;

pub use config::DeploymentMode;

pub use config::GatewayConfig;

pub use config::LocalModeConfig;

pub use config::RemoteModeConfig;

pub use config::TlsConfig;

pub use topology::EdgeType;

pub use topology::UnknownEdgeType;

pub use topology::cycle_detect;

pub use topology::Edge;

pub use topology::EdgeRepo;

pub use topology::EdgeRepoError;

pub use topology::NewEdge;

Modules

agent

Agent execution context carrying identity, PID, and metadata.

approval

Approval-related domain types shared across crates.

audit

Immutable, hash-chained audit entry for Agent Assembly governance events.

capability

Per-level capability types for fine-grained agent permission control.

config

Gateway deployment-mode configuration types (Epic 17, AAASM-1568).

dev_tool

Foundational types for the AI dev tool governance framework.

evaluators

Test-only PolicyEvaluator implementations.

identity

Stable identity types for agents and execution sessions.

llm

LLM provider and model identifiers shared across the workspace.

policy

Policy types and the PolicyEvaluator trait for governance decisions.

risk_tier

Risk tier classification for agents and policies.

storage

Storage trait abstraction for the Agent Assembly persistence layer.

time

Timestamp abstraction compatible with no_std environments.

topology

Agent-graph mesh topology types and repository trait (AAASM-985).

types

Stable, wire-shaped types shared by every storage driver (AAASM-2355).