AI Coding Tools Need the Same Policy Everywhere
Coding-agent governance only works when the same rules follow developers across every client and cloud task.
wade womersley – york based software engineer
Engineering notes from the sharp end
wade womersley – york based software engineer I write the way I work: direct, useful, and more interested in what holds up in production than what sounds clever on a slide.
267 published posts
Latest update
Latest post
Coding-agent governance only works when the same rules follow developers across every client and cloud task.
Archive
Euclid has found 31 exceptionally old quasars, including two from the universe’s first 670 million years, and the scale is hard to picture.
CloudFormation Express mode can cut deployment waits sharply, but configuration complete is not the same as ready for traffic.
.NET 8 and .NET 9 both reach end of support on November 10, so teams on either runtime should plan the move to .NET 10 now.
CodeQL 2.26 can trace untrusted values into JavaScript and TypeScript system prompts, turning one AI security risk into something code review can see.
GitHub Models closes on July 30, with brownouts first, so this is a migration deadline rather than a product update to read later.
PHP 8.6 Alpha 1 is not for production, but it is exactly when framework, library, and platform teams should begin compatibility testing.
GPT-5.6’s Sol, Terra, and Luna tiers make model choice a practical cost decision instead of a hunt for one universal default.
Android developer verification should be treated like release infrastructure, especially for teams distributing outside Google Play.
Jetpack Compose 1.11’s testing changes are worth planning for because timing assumptions can quietly break mobile test suites.