Tools
Free tools from real delivery work - run-it-yourself diagnostics that map your estate, and production-grade reference builds to clone, study and deploy.
Diagnostics
Run-it-yourself checks that map your estate and surface where it is quietly breaking - the entry point to a Delivery Blueprint.
Drop in a business-critical spreadsheet and see in seconds whether it is a risk - 17 checks covering key-person dependency, hidden errors, macros, external links and the point where it stops scaling. Runs entirely in your browser; nothing is uploaded. The engine is open source (MIT) and the exact bundle the site runs is published, hashed and provenance-attested, so you can audit what scored your workbook.
A headline Well-Architected score for your Azure estate, built from Azure Advisor signal and the Defender secure score. Deterministic, read-only, no sign-up, and nothing leaves your tenant - it runs from your own Azure CLI in about a minute. It is honest about its ceiling, and says so in every report it writes: 40+ of Microsoft's 59 checklist items cannot be read off a machine, and judging those is the full review.
Integration Health Check
In buildInventories what talks to what (and what silently doesn't) across common SaaS and database estates, and outputs a systems map.
Reference builds
Production-grade reference repos on GitHub - the senior judgement made concrete, not a slide.
A deploy-ready, opinionated reference build on Microsoft's official Azure AI Landing Zone Verified Module (AVM). Golden-path Terraform and tfvars for a secure, private-networked RAG workload in a regulated UK SME - the senior judgement layer the generic module leaves to you, with the money-pit resources flagged off by default so you deploy safe and cheap on day one.
A full-stack C#/.NET reference for regulated-grade AI on Azure: grounded, cited compliance Q&A over the public FCA Handbook. RAG over Azure AI Search (with a pgvector variant to prove both), and a tool-calling agent built on the Microsoft Agent Framework that looks up and cross-references specific provisions - every answer traced back to the source rule. This is what regulated AI actually needs, shipped end to end.
One container with the whole DevOps toolchain already in it - Terraform, Terragrunt, Azure CLI, Kubernetes, Ansible, PowerShell, .NET and the coding agents - so a new machine is productive after a pull, not a day of installs. Built natively for both Intel and Apple Silicon, rebuilt weekly so the tools stay current, and every image ships an SBOM and signed build provenance you can verify. It is honest about itself, too: a version tag is a point-in-time snapshot, and the digest is the only stable identity.
Agent tooling
Tooling around the coding agents themselves - adopting them safely across a team, and giving them your own notes and code to work from. Not marketplace skills; these stand on their own.
Recall
LiveHybrid search over your notes and code, spoken to your agent over MCP - BM25 and vector retrieval fused, so it finds what you meant when you cannot remember the words you used. Local-first: nothing leaves the machine. And it never degrades quietly - if half the search is not working, it says so, because a broken tool that looks healthy is worse than one that admits it.
AI-Engineering Starter Kit
In buildThe configuration, guardrail and workflow patterns for adopting agentic coding tools in a real team - written for a secure environment, genericised.
Fifteen minutes will tell you if this fits
Bring the problem - the stalled pilot, the systems that do not talk, the manual process eating your team. You will get a straight answer on whether it is sprint-shaped, roughly what it would cost, and when it could be running.
I reply within 24 hours