model-selection

Resolve the optimal Claude model for each agent by using a centralized runtime resolution entry point that evaluates configuration precedence and runtime overrides. This ensures deterministic orchestration by returning the selected model, its timeout, and the resolution source.

Configuration precedence for agent runtime overrides works by evaluating CLI overrides, environment variables, and config files sequentially before falling back to defaults. This resolution mechanism transparently reports the source to ensure consistent performance and resource use.

Yes, you can set per-agent model overrides for specific tasks like analyze, write, execute, validate, and fix. The system supports dynamic runtime overrides to ensure each agent uses the optimal Claude model for its designated operation.

The best way to centrally manage Claude model selection across multiple agents is to use a single entry point that applies deterministic defaults and safe fallbacks. This approach simplifies orchestration by transparently reporting the resolution source and timeout.

Agent runtime timeout resolution returns unexpected defaults when CLI overrides, environment variables, and config files fail to provide valid configurations. The system applies safe fallbacks to deterministic defaults to ensure consistent orchestration when higher-precedence sources are missing or invalid.

No, you do not need a config file to resolve deterministic model defaults for agent orchestration. The system applies safe fallbacks to defaults automatically, although config files provide a convenient way to manage per-agent overrides and timeout resolution.