ConTeXt generates .tua and .tuc files (utility cache) during each run. These files store document metadata between compilations and are not part of the source, so they should be ignored.
oh-my-opencode excludes the sisyphus working directory in their own repo, so it seems to be best practice. Also, the directory captures agent/runtime state, so it should not be shared.
Updated the days before closing stale PRs from 180 to 365. Due to internal priorities we need to allow more time for PR reviews.
Based on my reading, we need to increase `days-before-close` to ensure that any PR marked with the `stale` label currently benefits from the extension.
Once we are caught up on the backlog, we can revisit this setting.
Historically R users have used the .Renviron file for storing environmental variables such as secrets, credentials, API keys etc. However, I believe there are several compelling reasons for adding .env to the .gitignore too:
1. The dotenv R package was created to allow users to use .env instead of .Renviron. The package has a significant number of downloads. https://github.com/gaborcsardi/dotenv
2. .env files are standard in many other programming languages. It wouldn't be that out of the unexpected to think R users might end up with an .env file because a python user on their team contributed files to the repo, the R user started programming in nodejs and is used to working with .env file, or a repo is in the process of being translated from a language that uses .env to R.
3. LLMs are making programming accessible to people who may not have much previous experience handling credentials and security. One of the first steps many of these users have to do is store LLM provider API keys somewhere. I worry about a new R user adapting a python workflow or an R workflow that recommends using an .env file and then accidentally committing their LLM API keys to a public repo.
4. I can't think of why an R user would create an .env file for any other reason other than storing credentials and environmental variables. If, for whatever reason, they need to commit an .env file, they can remove that part from the template, but I think the risk of accidentally exposing sensitive credentials justifies adding .env files to the template and not commenting it out by default.
