Skills for Dlang?

[Lance Bachmeier]

On Sunday, 27 September 2026 at 11:22:24 UTC, Guillaume Piolat 
wrote:

> This is an attack vector for the LLM, but so it foreign code so 
> it can rely on the same trust system that already exist.

True, but LLM skills have an extremely bad record. This study 
found security issues in 37% of posted skills: 
https://snyk.io/blog/toxicskills-malicious-ai-agent-skills-clawhub/ That's likely due to a high correlation between AI use and not bothering to read code.