Code Repository
A version-controlled codebase — branches, commits, contributors, and CI/CD configuration that contains the product source code.
Why This Object Matters for AI
AI code completion and review analyze repository context; deployment automation depends on repository structure.
Engineering & Development Capacity Profile
Typical CMC levels for engineering & development in SaaS/Technology organizations.
CMC Dimension Scenarios
What each CMC level looks like specifically for Code Repository. Baseline level is highlighted.
Source code lives on individual developer laptops with no version control. 'Where's the latest code?' depends on which developer you ask. There's no central repository, no commit history, and no way to see who changed what. Deploying to production means copying files from someone's machine.
None — AI cannot analyze, review, or assist with code because no code repository exists in any accessible system.
Set up a version control system — create a Git repository hosted on GitHub, GitLab, or Bitbucket with all source code committed and a basic branching strategy.
A Git repository exists but practices are inconsistent. Some developers commit daily; others push weeks of work in one commit. Commit messages range from 'fix stuff' to paragraphs. Branch naming is random. The repo has no README, no documentation of architecture decisions, and no CI/CD configuration. The repository is a code dump, not a structured knowledge asset.
AI can access code files and generate basic suggestions, but cannot understand project structure, contribution patterns, or code ownership because the repository lacks consistent metadata, documentation, and organizational conventions.
Establish repository conventions — standardized branch naming, commit message format, required README with setup instructions, and a CODEOWNERS file mapping code areas to responsible engineers.
The code repository follows established conventions — standardized branch naming, commit message templates, CODEOWNERS file, and a README. CI/CD pipelines run on every push. Code is organized into logical directories. But repository-level metadata is limited — there's no architecture documentation, no dependency graph, and no machine-readable service boundary definitions.
AI can generate contextual code suggestions, run automated reviews against style guides, and navigate the codebase effectively. Cannot reason about architectural boundaries or service dependencies because these aren't documented in the repository.
Add structured repository metadata — architecture decision records (ADRs), machine-readable service boundary definitions, dependency manifests, and API contracts that document the codebase's structure beyond just the code files.
The code repository is a comprehensive knowledge asset. Architecture decision records explain design choices. Service boundary definitions are machine-readable. Dependency graphs are generated from manifests. An AI agent can query 'what services depend on the user authentication module and what were the architectural reasons for the current design?' and get a documented, accurate answer.
AI can generate code that respects architectural boundaries, suggest refactoring aligned with documented design principles, and assess the impact of changes on dependent services. Cannot yet autonomously evolve the architecture because design decisions require human strategic judgment.
Formalize the repository as a machine-readable software ontology — typed entity definitions for services, APIs, dependencies, and data flows with validated relationships that AI agents can query programmatically.
The code repository is modeled as a formal software ontology. Services, APIs, data flows, and dependencies are typed entities with validated relationships. An AI agent can ask 'generate a microservice that handles payment webhooks, following our established patterns for event-driven services, and respecting our data isolation boundaries' and produce architecturally consistent code.
AI can autonomously generate, review, and refactor code within the formal software model. Can produce architecturally consistent implementations for well-defined patterns. Human judgment is needed for novel architectural decisions and strategic technical direction.
Implement real-time repository intelligence — the software ontology auto-updates as code changes, keeping architectural documentation, dependency graphs, and service maps current without manual maintenance.
The code repository is a self-documenting software model. Every code change automatically updates architecture documentation, dependency graphs, and service maps. The repository generates its own structural knowledge from the code as it evolves. Architecture decisions are captured automatically from PR discussions and commit patterns.
Fully autonomous code intelligence. AI maintains, documents, and evolves the codebase with complete real-time understanding of structure, patterns, and architectural principles.
Ceiling of the CMC framework for this dimension.
Other Objects in Engineering & Development
Related business objects in the same function area.
Pull Request
EntityA code change proposal — diff, reviewers, comments, approvals, and merge status that gates code into production.
Test Suite
EntityA collection of automated tests — test cases, coverage metrics, and execution results that validate code quality.
Engineering Task
EntityA development work item — story, bug, or tech debt with estimates, assignee, and status that tracks engineering work.
Deployment
EntityA production release — version, changes, timing, rollback capability, and status that tracks code going live.
Technical Debt Item
EntityA documented code quality issue — description, impact, effort estimate, and priority for addressing accumulated debt.
What Can Your Organization Deploy?
Enter your context profile or request an assessment to see which capabilities your infrastructure supports.