COMPANION RESOURCE
The Silent Wiring Prompt Pack
Two kits for AI coding assistants. Free, and yours to pass on.
Build It Loud makes an AI coding assistant generate new event-driven integrations that cannot hide a dead edge. Audit It Loud turns the same three checks on code you already have.
Every prompt is tool-agnostic. Paste it into any assistant, fill the brackets, and hold the output to the acceptance criteria printed underneath it. Where a repo-aware agent can do meaningfully more, there is a variant for it, because the difference between reasoning about a check and actually running it is the entire point of the mutation layer.
What is in it
- Kit 1, Build It Loud. The flagship build-and-self-verify prompt, plus a quick variant for one or two edges.
- Kit 2, Audit It Loud. Reconstruct the actual topology, probe each edge for liveness, mutation-test the existing suite, rank the findings by risk, and harden a critical edge with the extended operators.
- A worked example. One real flow from a production system, carried through all three layers: the declaration, the lint output that caught it dead, the probe that proves it alive, and the mutations that prove the probe would notice.
- The three layers, plus a triage pass. Topology declaration, liveness verification, mutation testing, and the confidence pass that decides where to aim first.
- A shared vocabulary table, so the terms in the prompts line up with the terms in the book's field manual rather than drifting from them.
Take it
It is a single markdown file, about 4,300 words. Download it and open it wherever you already work, which is where you will be using it. There is nothing to sign up for and nothing is recorded when you take it.
The pack, as one file
Markdown, tool-agnostic, no attribution required. Send it to whoever on your team would use it.
Download the prompt packThe code, as files
Every fenced code example in the book and its field manual, 113 of them, extracted as individual files and indexed by the unit they came from. Python, TypeScript, YAML, shell, and the sample output from the checks themselves, because what a check prints is usually the point.
These are the book's examples, not a library. They use stand-in imports and assume a system that is not yours, so most will need editing before they run. They are here so you can read them properly, diff them against your own code, and hand an assistant a concrete shape instead of describing one from memory. Every file is byte-identical to the text in the book. Nothing has been tidied up or completed.
The code examples, as one archive
113 files, with an index and a short guide to where to start. Same terms as the pack: nothing to sign up for, nothing recorded.
Download the code examplesWhere it comes from
The pack is the practical end of a book about integration code that compiles, passes review, keeps its dashboards green, and moves nothing at all. The prompts here apply the three checks; the book is the argument for why the checks you already run cannot see this, and the account of the four I shipped into my own production system, which between them covered every shape the book describes.