Know your codebase is sound, auditable, and governable, whoever wrote it, human or AI.
AI makes delivery fast. It does not tell you the result is safe to own, to audit, or to buy. We give you an independent, evidence-based read of your code, and the guardrails that keep it that way. Assurance, not assertion.
Whether you build the software or you are about to buy it, the question is the same: can you trust what is in the repository.
Start with a free readiness scorecard, then a Readiness Assessment: a scored report, the risks an AI assistant will amplify in your code today, and a prioritized roadmap. Remediation installs the guardrails in place, so your engineers do not have to retrain to keep the standard.
Before you commit capital, an independent, committee-ready read of a target's codebase, legacy or AI-built: an asset, conditional, or liability verdict, ranked findings tied to the code, and the cost to bring it to a safe standard. Under NDA, on a fixed timeline.
Generative Specification is an open method, and you can use all of it without us. We offer services to learn it, put it in place and keep it running, including the governance and strategy layer. It works with the tools your team already uses, on a new project, an existing one, a takeover or a migration.
A letter for each property of your code and a short list of tasks, run on your own repository. Useful on its own.
A report you can show your board, with evidence from your code. The full amount counts toward the next step if you go on within 90 days.
We put the guardrails in your codebase, or teach your team to keep them. The scope comes from the assessment, so nothing is quoted blind.
"Everyone feels like a systems engineer capable of building things now, but the truth is it gets us into trouble, especially at publish time."
"A non-technical analyst on the team built an audit agent for their CRM from what he learned, documented it, and presented it to leadership. They called it complete."
The method behind the assessments is published, and you can inspect it. Write the specification as a contract, derive the code from it, and keep the guarantee outside the model. A buyer who cannot verify us can verify the method.
The paper tree (a base paper and focused derivatives), the models and equations with the status each has earned, the pre-registered experiments and their raw evidence, and the lessons written out. Open access, with a DOI.
Declared interest. genspec.dev is the open home of the method (CC BY 4.0), written by the same author who runs PragmaWorks. Its studies are small and author-run, so they are published to be re-run, with their limits. PragmaWorks sells services built on the method; the method does not depend on buying them.
Free tools for individual developers. Each one addresses a specific failure mode of AI-assisted development.
Portable quality gates, one per property, wired to standard tools: mutation, real coverage, complexity, duplication, dead code. Hand it to your AI and it cables the gates into your project straight from the spec. The ratchet only goes up.
Quality gates →Graph-powered hybrid search for your codebase. Vector + BM25 + path, fused with RRF. Semantic search, dependency traversal, and contextual file reads — the AI reads your codebase the way you do, not just grep.
GitHubThree-tier persistent memory for AI sessions. Buffer, Working, and Core layers with Ebbinghaus decay and an intelligence distillation layer. Cross-project knowledge that survives context boundaries.
GitHubSixteen hours. Your team. Your codebase. A different way of working.
Point AI at a codebase and you get plausible output fast, then pay for it in review and rework. The Forge teaches the discipline that makes the output correct enough to trust: spec-driven development, clean architecture, and directing AI against a spec instead of prompting into the dark. Sixteen hours, on your actual codebase, with your team.
Book a ConversationLong-form writing on specification, AI-assisted development, and what it means to engineer in an age of stateless readers.
ambientengineer.dev The Andon Method →Workshop inquiries, research questions, collaboration.