chore: agent operating rules and skills - #2006
Merged
Merged
Conversation
For this set of changes I'm following Thomas recomendations on eclipse-openvsx#1851, cherry-picking what I found useful for our use case, plus some additions. I'm also adding some skills to help agents migrate endpoints to tanstack query, basically since we're in the middle of a migration they might not have the full picture of what approach we're taking. I'm also adding some testing related skills, that pairs with eclipse-openvsx#2005 where I'm doing the groundwork for testing components. At the same time I'm creating symlinks so that we don't stick to claude code, but at the same time it works. Overall my approach with the `AGENTS.md` has been to keep it minimal, if we're asking the agent to do some explanation or planning it should not need to know how to crate a test or how we like to use react query. At the same time if we've already had some chatting with the agent the <system-reminder> block where the AGENTS.md is pasted sits already quite far away since it's not part of the system prompt. Skills on the other hand are invoked by the agent during the session so the rules/conventions should remain more fresh.
netomi
approved these changes
Jul 28, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
For this set of changes I'm following Thomas recomendations on #1851, cherry-picking what I found useful for our use case, plus some additions.
I'm also adding some skills to help agents migrate endpoints to tanstack query, basically since we're in the middle of a migration they might not have the full picture of what approach we're taking.
I'm also adding some testing related skills, that pairs with #2005 where I'm doing the groundwork for testing components.
At the same time I'm creating symlinks so that we don't stick to claude code, but at the same time it works.
Overall my approach with the
AGENTS.mdhas been to keep it minimal, if we're asking the agent to do some explanation or planning it should not need to know how to crate a test or how we like to use react query. At the same time if we've already had some chatting with the agent the block where the AGENTS.md is pasted sits already quite far away since it's not part of the system prompt. Skills on the other hand are invoked by the agent during the session so the rules/conventions should remain more fresh.