Hi Thomas,

On Sat, 25 Apr 2026 10:28:53 +0200 Thomas Koenig wrote:

> I think that patches which contain LLM-generated code need to receive
> special care from *reviewers*, as well (to be formalized how?)

To this aim, our corporate policy mandates inclusion of prompts in
commits containing LLM output (together with tool name and version).

Over time, this helped a lot developers here to build an effective
insight about the issues that might pass unnoticed.
Over time we also independently adopted a habit to isolate LLM output
into dedicated commits, that proved very useful during post-mortem and
unexpected-costs analysis (bugs, security incidents...).


Giacomo

Reply via email to