Expand description
Versioned secret-free coding settings and deterministic layer merging.
Re-exports§
pub use crate::mcp_config::McpArgumentResourceSettings;pub use crate::mcp_config::McpConcurrencySettings;pub use crate::mcp_config::McpLifecycleSettings;pub use crate::mcp_config::McpLimitsSettings;pub use crate::mcp_config::McpReconnectSettings;pub use crate::mcp_config::McpRetrySafetySettings;pub use crate::mcp_config::McpServerSettings;pub use crate::mcp_config::McpToolDeclarationSettings;pub use crate::mcp_config::McpToolSettings;pub use crate::mcp_config::McpTransportSettings;
Structs§
- Client
WebSearch Settings - Client web-search backend controls resolved without reading credentials.
- Client
WebSearch Settings Layer - Sparse client web-search backend overlay.
- Coding
Settings - Versioned, secret-free resolved settings snapshot.
- Model
Capabilities Config - Explicit, fail-closed capabilities for one configured model and endpoint.
- Model
Definition - One model advertised by a configured provider.
- Model
Reasoning Config - Strict model-level reasoning profile and provider wire overrides.
- Provider
Config - Sparse OpenAI-compatible provider connection and model catalog.
- Provider
Value Resolver - Resolves provider values from environment templates and cached shell commands.
- Providers
Config - Provider map loaded from a standalone
providers.jsonfile. - Providers
Config Load - Soft-failing provider configuration load result.
- Resource
Settings - Declarative resource discovery settings.
- Resource
Settings Layer - Sparse resource overlay.
- Settings
Layer - Sparse one-layer settings overlay.
- TuiSettings
- Nested TUI settings merged independently by layer.
- TuiSettings
Layer - Sparse TUI overlay.
- WebFetch
Cache Settings - Bounded in-memory cache controls for normalized fetch results.
- WebFetch
Cache Settings Layer - Sparse normalized-result cache overlay.
- WebFetch
Settings - Secret-free resolved web-fetch settings.
- WebFetch
Settings Layer - Sparse web-fetch settings overlay.
- WebSearch
Location Settings - Approximate location fields used by hosted web search.
- WebSearch
Location Settings Layer - Sparse approximate-location overlay.
- WebSearch
Settings - Secret-free resolved web-search settings.
- WebSearch
Settings Layer - Sparse web-search settings overlay.
Enums§
- Hosted
Tool Capability - Provider-hosted capability accepted in custom model configuration.
- Project
Trust Default - Behavior when project-local configuration exists without a saved decision.
- Providers
Config Load Error - Safe classification for a provider configuration load failure.
- WebFetch
Backend - Client web-fetch backend selected by the embedding product.
- WebSearch
Client Backend - Client web-search backend selected by the embedding product.
- WebSearch
Route Preference - Route selected for web search when a real client backend is available.
Constants§
- CODING_
SETTINGS_ SCHEMA_ VERSION - Current persisted coding settings schema.
- DEFAULT_
WEB_ SEARCH_ API_ KEY_ ENVIRONMENT - Default environment variable consulted by an embedding product for Tavily credentials.
- DEFAULT_
WEB_ SEARCH_ TIMEOUT_ MILLIS - Default client web-search request timeout.
- MAX_
PROVIDERS_ FILE_ BYTES - Maximum accepted size of one provider configuration file.
- MAX_
SETTINGS_ FILE_ BYTES - Maximum encoded settings file bytes.
- MAX_
WEB_ SEARCH_ API_ KEY_ ENVIRONMENT_ BYTES - Maximum bytes in one configured API-key environment variable name.
Functions§
- load_
providers_ file - Loads a strict, bounded provider configuration without propagating file errors.
- load_
settings_ file - Loads one optional strict JSON settings layer.
- merge_
settings - Deterministically resolves defaults < global < trusted project < env < CLI.
- persist_
global_ model_ settings - Atomically updates only the global model and reasoning defaults.