Expand description
Compatibility package that re-exports the primary awaken umbrella crate.
New code should depend on the published awaken package directly. This
package exists so existing awaken-agent users can move to the new release
line without changing import paths.
Modules§
- agent
- builder
- Fluent builder API for constructing
AgentRuntime. - context
- Context management: compaction, summarization, truncation, and plugin state.
- contract
- engine
- Engine layer: genai-backed LLM executor and type conversion.
- execution
- Tool execution concerns: executors.
- ext_
generative_ ui - ext_mcp
- Model Context Protocol (MCP) client integration for external tool servers.
- ext_
observability - LLM telemetry plugin aligned with OpenTelemetry GenAI Semantic Conventions.
- ext_
permission - Permission extension for the awaken agent framework.
- ext_
reminder - Reminder extension for the awaken agent framework.
- ext_
skills - Skill subsystem for awaken.
- extensions
- Bundled extensions for the awaken runtime.
- loop_
runner - Minimal sequential agent loop driven by state machines.
- model
- phase
- plugins
- policies
- Stop condition policy system and built-in policies.
- prelude
- One-import convenience module for awaken.
- registry
- Registry traits, in-memory implementations, and agent resolution.
- registry_
spec - Serializable agent definition — pure data, no trait objects.
- runtime
- server
- HTTP server layer for the awaken agent framework.
- state
- Combined state types from both the contract and runtime layers.
- stores
- Storage backend implementations for the awaken framework.
Structs§
- Agent
Runtime - Top-level agent runtime. Manages runs across threads.
- Agent
Runtime Builder - Fluent API for constructing an
AgentRuntime. - Agent
Spec - Serializable agent definition referencing registries by ID.
- Cancellation
Token - A cooperative cancellation token.
- Commit
Event - Execution
Env - Per-run execution environment.
- Failed
Scheduled Actions - Mutation
Batch - A batch of state mutation operations.
- Pending
Scheduled Actions - Persisted
State - Phase
Context - Execution context passed to phase hooks and action handlers.
- Phase
Runtime - Plugin
Descriptor - Plugin
Registrar - Redacted
String - String wrapper whose
Debug/Displayimplementations never reveal the underlying value, and whose buffer is zeroized on drop. - Resolved
Agent - A fully resolved agent: all capabilities + plugin environment, ready to run.
- RunRequest
- Unified request for starting or resuming a run.
- Snapshot
- State
Command - A command that carries state mutations, scheduled actions, and effects.
- State
KeyOptions - State
Map - State
Store - Typed
Effect
Enums§
- Build
Error - Error returned when the builder cannot construct the runtime.
- Json
Value - Represents any valid JSON value.
- KeyScope
- Lifetime scope for a state key.
- Merge
Strategy - Parallel merge strategy for a state key.
- Phase
- Lifecycle phase within an agent run.
- Runtime
Error - Runtime-specific errors that wrap
StateErrorand add variants for agent resolution and run management. - State
Error - Unknown
KeyPolicy
Constants§
Traits§
- Agent
Resolver - Resolves an agent by ID, producing a ready-to-execute config + environment.
- Commit
Hook - Effect
Spec - Phase
Hook - Plugin
- Plugin
Config Key - Typed plugin configuration key.
- Scheduled
Action Spec - State
Key - Tool
Gate Hook - Typed
Effect Handler - Typed
Scheduled Action Handler - Typed
Tool - A strongly-typed tool trait that derives its descriptor schema from
Args.
Functions§
- generate_
tool_ schema - Generate a JSON Schema for
Tthat is suitable for LLM tool calling. - sanitize_
for_ llm - Sanitize a JSON Schema value in-place to be LLM-friendly.
- validate_
against_ schema - Validate
argsagainst a JSON Schema, returning an error with joined messages on failure.