The manifesto

Seven letters. Seven breaths.

INSPIRE is not an acronym stretched over a process. It is seven properties that allow a software product and its underlying knowledge to emerge, evolve and remain coherent in the agentic era.

I

Intelligent

Native to artificial intelligence.

INSPIRE is not a pre-AI process with agents bolted on. It starts from what AI makes possible, orchestrating agents at scale, and puts intelligence into every stage: skills that interview, agents that build and oversee, and humans who bring intent, context and judgment.

N

Navigable

Context remains discoverable.

A navigable system connects intent, specifications, decisions, implementation and verification into a coherent knowledge graph. Humans and agents can discover context, trace decisions and evolve the system without losing understanding.

S

Spec-driven

Intent becomes the source of truth.

The specification captures what the system should be, while code becomes an expression of that intent. Implementation can change, but the knowledge behind it remains durable, explicit and verifiable.

P

Prototypical

Explore before committing.

Prototypes are not early versions of the product; they are disposable instruments for learning. They make uncertainty visible, allowing ideas to evolve before becoming part of the persistent knowledge layer.

I

Iterative

Progress comes from refinement.

The first version is not expected to be right — it is expected to reveal what is missing. Each iteration reduces uncertainty, revisits previous decisions and moves the system closer to what it should become.

R

Robust

Vibe-coding speed, engineering-grade results.

Working with agents should feel as fast and flexible as vibe coding: prototype freely, change direction within a single Breath. What changes is what's underneath. Coherence can't rely on discipline or memory, so it is protected by design: guardrails, skills and automated checks turn the specification's rules into constraints that guide agents, and every change has to pass them before it lands. The result holds up under change, scale and time.

E

Executable

The specification executes like a senior engineer.

The knowledge graph is complete and precise enough that agents emanate production code from it the way a senior engineer would: honouring the architecture decisions, implementing the contracts, following the stack's conventions, writing the tests first and getting the work reviewed before it lands. When an implementation drifts, it is emanated afresh from the spec instead of patched.

Principles are only as good as the practice that carries them.

The methodology puts these seven into practice, the skills make them repeatable, and the factory makes them hold at scale.

Stop sprinting.
Start breathing.