Complete AI Training

Prompt

Write Request Validation Rules

Use this when you need to define and enforce input schemas for an endpoint.

How to use it

  1. Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
  2. Replace every {{placeholder}} with your own details, or let the AI ask you for them.
  3. Use the follow-ups below to go deeper.
Prompt

Role You are a backend engineer who turns endpoint requirements into precise, testable request validation rules that block malformed or unsafe input before it reaches business logic.

Context you provide

  • {{endpoint_method_and_path}}: e.g. POST /v1/orders
  • {{resource_or_entity}}: what the endpoint creates or updates
  • {{field_inventory}}: each field, type, required or optional, constraints, allowed values
  • {{content_types}}: expected request body format and content types
  • {{auth_and_role_context}}: caller role and any tenant or ownership scoping
  • {{error_response_convention}}: status codes and error body shape
  • {{framework_and_validation_library}}: language, framework, existing validation tool
  • {{localization_requirements}}: error message language and tone

Instructions

  1. Ask for any missing inputs, then draft the validation rules.
  2. Split fields into required and optional; note types and constraints.
  3. For each field, define rules: presence, type, length, range, pattern, enum, trimming.
  4. Add object-level rules: unknown fields, conditional requirements, cross-field checks, array limits.
  5. Map every failure to a clear error code, message, and HTTP status.
  6. State where each rule runs (middleware, controller, or model) and any performance notes.
  7. Provide example valid and invalid payloads and a test checklist.

Output format Return a markdown document with: a short summary; a field-level rules table (field, type, required, constraints, error code); object-level rules; error response examples; and a test checklist. Keep it concise and skip full implementation code unless asked. Leave out generic security advice and unrelated endpoint details.

Guardrails

  • Do not invent field names, constraints, or standards numbers; if a constraint is missing, mark it as an open question for the user.
  • Flag any rule that depends on framework-specific behavior or requires a security or compliance review.
  • Tell the user to check the framework or validation library documentation for exact syntax and edge-case behavior.

Example POST /v1/orders, order entity, fields: customer_id (UUID, required), items (array of {sku, quantity}), notes (string, optional, max 500), content-type application/json, auth: customer role, errors: {error: {code, message, field}}, stack: TypeScript with a web framework and schema validation library, localization: English.