Configuration Reference

Canon operates via governed packets and standard repository conventions rather than a centralized, monolithic configuration file. Its "configuration" is defined through its modes and the structures it creates within the repository.

Governing Context

Canon uses semantic governance to maintain architectural integrity. To configure Canon's behavior, you define the constraints and rules directly in the relevant domain models or instructions:

• Mode Configuration: Canon's CLI allows you to execute different modes (e.g., pr-review, init). The configuration for these modes is often derived from the target context and explicitly provided arguments.
• Project Memory: Instead of static JSON configurations, Canon treats your tech-docs/ and specs/ directories as a living configuration base. AI systems read these structured documents to align with your project's Domain Language.
• Evidence Requirements: The rules for what constitutes acceptable evidence are embedded in the mode templates, meaning configuration is inherently tied to the workflow you select.

Integration Adapters

Canon supports Governance Adapters to integrate with other tools (like Boundline or GitHub). Configuration for these adapters typically involves:

• Ensuring the adapter binary is in your PATH.
• Setting up environment variables for remote endpoints (if using hosted solutions).
• Defining the target repository and branches using standard Git workflows.

For detailed configuration of specific commands, see the CLI Reference.