Crate brainwires_autonomy 

brainwires-autonomy

Autonomous agent operations — self-improvement, Git workflows, and human-out-of-loop execution for the Brainwires Agent Framework.

Feature flags

Feature	Description
self-improve	Self-improvement controller and strategies
eval-driven	Eval-driven feedback loop (requires brainwires-eval)
supervisor	Agent supervisor with health monitoring
attention	Attention mechanism with RAG integration
parallel	Parallel coordinator with optional MDAP
training	Autonomous training loop
git-workflow	Automated Git workflow pipeline
webhook	Webhook server for Git forge events
full	All features enabled

Re-exports

pub use config::AutonomyConfig;

pub use error::AutonomyError;

pub use metrics::SessionMetrics;

pub use metrics::SessionReport;

pub use safety::ApprovalPolicy;

pub use safety::AutonomousOperation;

pub use safety::SafetyGuard;

Modules

agent_ops

Agent operations — supervisor, health monitoring, attention, parallel coordination, hibernation data types, and autonomous training.

config

Configuration types for autonomous operations.

error

Error types for the autonomy crate.

metrics

Metrics and reporting for autonomous improvement sessions.

safety

Safety mechanisms for autonomous operations.