Spec-driven development
Definition
Spec-driven development builds AI systems (agents, pipelines, tools) from explicit specifications: requirements, output formats, allowed actions, and constraints. Specs are retrieved and used at runtime (e.g. in RDD) so behavior stays aligned with intent.
It is especially useful for agents and RDD: instead of encoding all rules in weights or prompts, you maintain specs (e.g. in docs or a knowledge base) and retrieve them at runtime. Fits regulated domains and teams that want behavior to be auditable and updatable without retraining.
How it works
You write specs (natural language, schemas, or structured rules) and index them for retrieval (e.g. in a vector store or structured repo). At runtime, the task (and optionally the current state) is used to retrieve relevant spec fragments. The model or agent decides (e.g. next step, allowed actions) and generates (output, tool call) with the spec in context. Validate checks the output against the spec (e.g. schema, rules); if validation fails, you can retry or surface an error. This keeps generation and decisions aligned with the spec without baking everything into prompt engineering or fine-tuning.
Use cases
Spec-driven development fits when behavior must stay aligned with retrievable requirements (RDD, compliance, or safety).
- Building agents that retrieve and follow specs (e.g. RDD pattern)
- Enforcing output format and constraints (JSON, allowed actions)
- Regulated or safety-critical flows where behavior must match requirements
External documentation
- LangChain – Structured output — Enforcing output format from LLMs
- OpenAI – Structured outputs