FakerForge

Documentation

Faker Forge Docs

Complete user guides for Data Generation, API Mocking, and MCP integration, organized for both developers and LLM-driven workflows.

Menu

MCP

MCP Workflows

Common end-to-end agent workflows for generation and mocking.

MCP Workflows

These workflows show practical, repeatable agent-driven operations.

Workflow 1: New Codebase Generation

  1. Derive codebase identifier from git remote.
  2. Call generate-fake-data with full schema.
  3. Poll until completion.
  4. Retrieve generated result references.

Workflow 2: Evolving Existing Schema

  1. Call get-codebase-schema with codebase identifier.
  2. Add newly introduced tables.
  3. Add/remove changed columns.
  4. Regenerate fake data.

Workflow 3: Shape Realistic Data

  1. Call get-schema-overview to see tables, profiles and relationships.
  2. Call get-column-profiles for the table you want to improve.
  3. Call update-column-profile for precise changes, or reprofile-tables for broad ones.
  4. Call preview-rows and show the user the sample.
  5. Call generate-fake-data to apply the changes, then export-generated-data.

Workflow 4: Fix Relationships

  1. Call get-schema-overview and look for *_id columns without a relationship.
  2. Call read-schema-source for those tables to confirm the intended foreign key.
  3. Call manage-relationship with action: add.
  4. For a wrong relationship, describe the change, wait for the user to confirm, then call manage-relationship with confirmed: true.
  5. Use set-parent-tables and set-row-counts to shape ratios, then regenerate.

Workflow 5: Mock API Bootstrap

  1. Call create-api-mock with domain prompt.
  2. Review generated endpoints.
  3. Call get-api-mock for exact endpoint/scenario details.
  4. Iterate prompt for additional coverage.

Workflow 6: Extend Existing Mock

  1. Call list-mocks.
  2. Pick mock by ID.
  3. Call create-api-mock with mock_id and prompt for only new endpoints.
  4. Validate responses in client tests.

Workflow 7: Safety-First Operation Pattern

  1. Start with read-only tools (list/get).
  2. Validate context and identifiers.
  3. Use mutation tools intentionally.
  4. Re-read state and confirm expected changes.

Team Conventions

  • Store shared codebase identifiers in project docs.
  • Keep MCP prompts deterministic and explicit.
  • Prefer incremental mutations over large one-shot changes.