Codebase Atlas

    1

    Generate evidence-backed architecture diagrams and code walkthroughs to map and document software repositories.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Codebase Atlas

    Example session with this skill installed

    Generate a workflow diagram for our user registration process in the services/auth directory and explain the main components.

    • Read your context and instructions
    • Compiled the codebase atlas

    I have analyzed the `services/auth` directory and its related entry points. Based on the source code in `registration.go`, `validator.go`, and the `user` database schema, here is the registration workflow. ## User Registration Workflow This diagram traces a registration request from the API gateway through validation to persistence in the Postgres store. ```mermaid graph TD Client[Client Browser/App] -->|POST /api/register| Gateway[API Gateway] Gateway -->|Forward Request| AuthSvc[Auth Service: registration.go] AuthSvc -->|Validate Payload| Val[Validation Logic: validator.go] Val -->|Check Complexity| PWD[Password Hasher: bcrypt] AuthSvc -->|Check Duplicate| DB[(User Store: Postgres)] AuthSvc -->|Write Record| DB AuthSvc -->|Emit Event| EventBus[Redis Event Bus] style AuthSvc fill:#f9f,stroke:#333,stroke-width:2px style DB fill:#bbf,stroke:#333,stroke-width:2px ``` ### Component Responsibilities - **Registration Logic (`services/auth/registration.go`)**: Orchestrates the flow, handles hashing, and manages the database transaction. - **Validation (`services/auth/validator.go`)**: Enforces input constraints and sanitizes data before processing. - **User Store (`db/schema.sql`)**: Confirmed via migrations to use a unique constraint on the email field. - **Event Bus**: Inferred from the `Close()` method in `pubsub.go`, used for downstream notification of new user accounts. ### Evidence & Assumptions - **CONFIRMED**: Database writes and password hashing are directly visible in `registration.go`. - **INFERRED**: The connection to Redis for the Event Bus is implied by the `config/queue.yaml` settings but the specific emitter call was not found in the primary registration path. **Next steps** 1. Review the `services/auth/registration.go` file to confirm the Redis emitter implementation. 2. I can save this diagram to `docs/architecture/registration_workflow.md` if you need persistent documentation. 3. Ask for a **Deep Dive** into the `Password Hasher` if you need to see specific security parameters.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Onboarding to a new codebase usually involves hours of reading imports and chasing function calls just to understand the data flow. This skill automates that discovery process, producing evidence-backed architecture diagrams and walkthroughs directly from your source. ## What it does - **Visualizes Architecture**: Generates Mermaid diagrams for systems, components, and workflows. - **Maps Workflows**: Traces execution paths and call sequences across multiple files. - **Documents Databases**: Creates entity relationship diagrams by inspecting schemas and migrations. - **Identifies Dependencies**: Maps interactions between subsystems and external services. - **Maintains Documentation**: Updates existing architecture files in `docs/architecture/` with version tracking. ## How it works 1. **Scope definition**: The agent identifies the subsystem or workflow you want to visualize and selects an investigation depth. 2. **Static exploration**: It inspects manifests, entry points, and configurations to build a conceptual map without running your code. 3. **Evidence collection**: Every node and relationship is categorized as confirmed or inferred based on source code analysis. 4. **Diagram generation**: The agent produces GitHub-compatible Mermaid code accompanied by a plain-language explanation and file citations. ## Frameworks & tools This skill is language-agnostic and works with any standard software repository. It thrives in environments with `README.md` files, standard manifests (e.g., `package.json`, `go.mod`, `requirements.txt`), and structured schemas. ## Why this beats prompting it yourself Generic prompts often hallucinate relationships based on file names or outdated documentation. This skill enforces a strict evidence-based procedure, distinguishing between confirmed source logic and inferred patterns while providing direct file references you can verify. ## Use cases - Map the data flow of a complex authentication or checkout workflow. - Generate a high-level system overview for onboarding new developers. - Visualize database relationships and foreign key constraints for a specific module. - Create sequence diagrams to debug race conditions or interaction order. ## Known limitations Does not execute code, installation scripts, or migrations. Analysis is limited to static source code and available documentation. Verification is only as accurate as the accessible files.

    Requires static access to source files. Cannot run scripts or migrations to determine runtime behavior. Accuracy depends on the clarity of code and manifests.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 7 days ago

    • Free with an Agensi account

    Requires

    Node.js

    Permissions

    Read Files
    Write Files
    Listed7 days ago

    What's inside

    Frequently Asked Questions

    The skill uses static analysis to trace execution paths across files, generating Mermaid.js code for system overviews, sequence diagrams, and ERDs. It specifically solves the 'onboarding tax' by automating the manual mapping of unfamiliar data flows.

    Yes, the skill is language-agnostic. It builds its mental map by analyzing universal patterns like manifests, imports, and directory structures, making it equally effective for TypeScript, Go, Python, and other common languages.

    Unlike generic prompts, this skill performs an 'evidence collection' step where every link in a diagram is tagged as either confirmed by code or inferred. It provides direct file citations for every relationship so you can audit the accuracy immediately.

    Codebase Atlas can automatically create or update files within your `docs/architecture/` folder. It includes version tracking to help you maintain a living history of your system design as your code evolves.

    The skill performs static code analysis only; it does not execute your code, run install scripts, or access live databases. This ensures your environment remains secure and side-effect free during the mapping process.

    You receive the full skill package, including the specialized investigation logic, Mermaid generation templates, and the evidence-based verification workflow designed for professional AI agents.

    New in Technical Documentation