harness-settings

Centralize provider-specific harness settings by storing per-provider source files under .harness/src/settings and applying them to native config locations like .claude/settings.json and .codex/config.toml. This enforces consistent configuration across multiple AI coding providers.

Manage multi-provider AI coding configurations by using harness plan and apply cycles to generate provider-specific files like .github/copilot-settings.json. This ensures teams maintain consistent provider configurations across CI/CD pipelines and multiple environments.

Yes, you can scaffold provider-specific config files individually using CLI commands like harness add settings claude, harness add settings codex, or harness add settings copilot. Each command generates the corresponding source file under .harness/src/settings.

The harness plan and apply workflow previews configuration changes and then writes provider artifacts to their native config locations. It maps source files like claude.json to outputs like .claude/settings.json, ensuring consistent application across providers.

Yes, you need separate per-provider source files such as claude.json, codex.json, and copilot.json stored under .harness/src/settings. Each file targets a specific provider and maps to its respective native configuration format during the apply cycle.

Harness settings may not apply correctly if the per-provider source files under .harness/src/settings are missing or misconfigured. Ensure you run harness plan to preview changes and harness apply to write the provider artifacts to their native config locations.