timsonner OpenCode Rules Configuration
Define custom instructions and behaviors for the OpenCode AI agent using rules.
Overview
Rules allow you to customize how OpenCode behaves in your project by defining specific instructions that the AI agent should follow. These can include coding conventions, project-specific guidelines, or behavioral constraints.
Creating Rules
Rules are typically defined in your project's configuration files or through special command interfaces. They help tailor OpenCode's behavior to your specific project needs and team practices.
Types of Rules
OpenCode supports various types of rules that can influence:
Coding Standards
- Language-specific formatting guidelines
- Naming conventions
- Code organization preferences
- Commenting standards
Behavioral Guidelines
- How the agent should approach problem-solving
- Preferences for certain implementation patterns
- Constraints on what the agent can or cannot do
- Interaction preferences with users
Project-Specific Instructions
- Framework-specific guidelines
- Architecture preferences
- Dependency management rules
- Testing strategies
Applying Rules
Rules can be applied through:
- Configuration files in your project
- Special command interfaces in OpenCode
- Agent-specific instructions
- Skill definitions that encode rules
Best Practices
When defining rules for OpenCode:
- Be specific and clear about what you want
- Focus on behaviors that will improve consistency
- Consider your team's existing practices
- Start with a few key rules and iterate
- Review and update rules as your project evolves
Example Rule Applications
While the specific syntax for defining rules may vary based on how you configure OpenCode, common applications include:
- Enforcing specific code formatting styles
- Requiring certain types of comments or documentation
- Specifying preferred architectural patterns
- Setting constraints on file modifications
- Defining how the agent should ask clarifying questions
Relationship to Other Configuration
Rules work alongside other OpenCode configuration elements:
- Tool permissions control what actions the agent can take
- Agent definitions specify capabilities and behaviors
- Skills provide domain-specific knowledge
- Models determine the underlying LLM capabilities
By defining clear rules, you can help ensure that OpenCode behaves in ways that align with your project's standards and your team's preferences.