trellis-meta

To customize Trellis workflows safely, you modify local .trellis architecture files rather than upstream sources. This approach preserves existing project conventions while routing changes to the correct local configuration entry points for tasks, context loading, and integrations.

The .trellis configuration architecture defines local project behavior through files, platform directories, hooks, agents, skills, commands, and workflows. It allows teams to adapt AI agent operations and platform integrations to fit specific project conventions and requirements.

To modify hooks and agents in a Trellis project, you target the local .trellis configuration files. The system guides you to the correct customization points for these components, ensuring changes are applied locally without affecting generated state or upstream sources.

Yes, Trellis supports local project configuration for custom AI agent workflows. It provides a structured architecture where platform directories, hooks, and skills can be customized locally to adapt agent behavior according to your team's specific project conventions.

You should avoid local Trellis customizations when edits require modifying upstream sources or generated state files. Changes must be confined to local .trellis configuration boundaries to prevent incorrect edits and preserve the integrity of the project architecture.