John Locke Describes Multi-Repository Workflow for AI-Assisted Development

Managing AI coding assistants across multiple repositories can become difficult when features span deployment pipelines, monitoring systems, automation tools, and application code. In a blog post on Freelock, founder and lead developer John Locke described a workflow called "Argo" that reorganises related repositories beneath a shared parent project to provide coding agents with broader operational context. The approach is intended to help AI assistants work across interconnected systems while preserving repository-level focus.

According to Locke, the workflow places repositories under a common parent directory that contains shared documentation, architectural guidance, development principles, and agent instructions. Tools such as CLAUDE.md allow coding assistants to inherit broader project context from parent directories while repository-specific instruction files continue to define implementation details and local conventions. Locke writes that the structure emerged in response to the challenge of coordinating changes across multiple internal tools and codebases.

The article also introduces "context markers", visual indicators added to agent responses to confirm that relevant instruction files have been loaded. Different repositories are assigned distinct symbols, allowing developers to verify whether both parent-level and repository-specific guidance is active. Locke presents the workflow as an extension of project documentation practices for environments in which both developers and AI coding assistants follow the same operational guidance, though the post does not provide quantitative measures of its impact.