Module scattered_types
Expand description
Config types that were originally defined in their home modules (agent, channels, tools, trust) but are needed by the config schema. Moved here to break circular dependencies.
Structs§
- Auto
Classify Config - Browser
Delegate Config - Channel
Precheck Config - Per-channel reply-intent precheck configuration.
- Clawd
Talk Config - Context
Compression Config - Email
Config - EmailO
Auth2 Config - OAuth2 settings for IMAP email authentication (XOAUTH2 / RFC 8628). Populate these to use OAuth2 bearer tokens instead of plain-password LOGIN. All major providers (Microsoft, Google) support this flow via device code.
- Eval
Config - Eval
Harness Config - Configuration for the agent evaluation harness (
[eval]), surfaced via thezeroclaw evalcommand. Distinct from[agent.eval], which is the in-loop response-quality scorer. - Gmail
Push Config - History
Pruner Config - Native
Thinking Params - Parameters for native extended thinking support.
- Thinking
Config - Configuration for thinking/reasoning level control.
- Trust
Config - Voice
Call Config
Enums§
- Thinking
Level - How deeply the model should reason for a given message.
- Voice
Provider - Which telephony model_provider to use.
Constants§
- MAX_
BUDGET_ TOKENS - MIN_
BUDGET_ TOKENS - Anthropic’s documented minimum for extended-thinking
budget_tokens. Requests below this are rejected with 400 by the provider; clamping at resolution time gives a clearer error site than the first API call.