AI agents don't read the AI policy #1983
johnbillion
started this conversation in
General
Replies: 2 comments 3 replies
-
|
Hi @johnbillion, thanks for opening this. I'm of two minds here:
I think my happy middle ground here would be to keep the AI policy single-sourced, but have AGENTS.md reference it with some additional language like "you must stop and show this link to the user; they must review the AI policy before contributing." Do you think something like that would work? |
Beta Was this translation helpful? Give feedback.
3 replies
This comment was marked as spam.
This comment was marked as spam.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
There seems to be two problems that prevent AI coding agents from reading the AI policy before starting work:
@CONTRIBUTING.mdinto an AGENTS.md and a CLAUDE.md file still results in the agents ignoring the instruction to read the AI policy. This seems to be because it requires a fetch from another repo rather than reading the policy from a local file.I've tested this with Codex and Claude Code and both consistently ignore the instruction unless I copy the policy into a local AI_POLICY.md file and then point to it in CONTRIBUTING.md. After doing this, both seem to consistently read the policy and respect it, resulting in a challenge such as the following:
Would you be open to copying the policy into a local file? It makes maintenance a touch more laborious but without doing so agents just ignore the instruction.
Beta Was this translation helpful? Give feedback.
All reactions