Feature hasn't been suggested before.
Describe the enhancement you want to request
AI agents seem disproportionately likely to reach for git commit --no-verify when hooks fail.
I understand why. They optimize for task completion. But I am a bit tired of catching this after the fact.
I know this can already be configured via permissions, but I wonder if the default behavior should be slightly safer here.
--no-verify intentionally bypasses repository-defined safety and quality constraints. For AI agents, this feels closer to a high-risk action than a normal git command.
Could git commit --no-verify (and git push --no-verify) default to ask?
Same argument may apply to a few other “bypass” actions (--force, publish, etc.), but --no-verify is the one I keep seeing in practice.
Feature hasn't been suggested before.
Describe the enhancement you want to request
AI agents seem disproportionately likely to reach for git commit --no-verify when hooks fail.
I understand why. They optimize for task completion. But I am a bit tired of catching this after the fact.
I know this can already be configured via permissions, but I wonder if the default behavior should be slightly safer here.
--no-verify intentionally bypasses repository-defined safety and quality constraints. For AI agents, this feels closer to a high-risk action than a normal git command.
Could git commit --no-verify (and git push --no-verify) default to ask?
Same argument may apply to a few other “bypass” actions (--force, publish, etc.), but --no-verify is the one I keep seeing in practice.