Posted by fg137 15 hours ago
otherwise a simple symlink from AGENTS.md -> CLAUDE.md works well enough
disclaimer, I only use open weight models and open source harnesses so have no stake in this either way, other than I support devs who do use claude (for now) and the symlink solution has worked fine for us
Sure its small, but it adds up and is just annoying overhead for most teams.
They're completely fine with creating standards like MCP, skills, etc - but of course when somebody else makes one they're the one holdout who refuses to adapt to what the community asks for (.agents folder, AGENTS.md, etc).
Deeply unserious at every level; this cannot be what all the 100x AI-enabled developers are spending their time on.
Most of the agents.md and what people use it for / write into is does, in fact, not make a difference.
Now, sure, this study is a bit old for LLM standards - as everything beyond the current month is - but
a) I haven't seen any tangible evidence to the contrary and
b) Since the basic inner workings of LLMs haven't changed I'd be sceptical of this not still applying.
I think one major side effect of LLMs moving so fast is that best practices and how to use this tool is very much not catching up as fast.
No one knows what is best and what actually makes a difference, doubly so because LLMs are / very / hard to quantify - even benchmarks themselves are very rough estimations.
People do, in fact, use stuff which makes no difference all the times.
> We conclude that while context files are useful for specifying non-standard coding practices, any attempts to improve performance should be rigorously evaluated before deployment.
The purpose of AGENTS.md is not to improve "coding performance" as the study looked at, it's to give an agent practical instructions that are useful to your specific workflow. For example you want it to use a certain format or specific tools for your project. This is stuff that can't be learned during training and must be loaded into the agent's context at the project level.
They also recommend starting with an /init command, which is also something the study very specifically called out as "having a marginal negative effect"
And from personal experience, I can only confirm that many people seem to see this as the main purpose of agents/claude md files - a persistent architectural overview of your project.
Quick Edit: My point simply being that I think it's understandable if some people don't understand the big deal about these files because they've had a drastically different experience than other people - anthropic themselves recommend apparently totally ineffective practices on their website, and the starter tool present in many harnesses seems to even have a (marginal) negative effect.
I suspect eventually this will move to fine-tuning, there are multi-tenant LoRAs on the horizon that will let us keep paying per-token on FTs
I’ll also include entries in the .github dir in project repos doing same thing in case anyone working on it happens to use GitHub copilot via vscode will also pick up on the AGENTS.md as well as any skills I might have cooked up for the repo.
(Though it’s mostly because I don’t trust team members to read the docs and the skills I made are meant to guide following standards for the project - and this approach almost incepts the standards for anyone not paying attention)
Both techniques have worked (and been officially documented by Anthropic) for as long as I can remember.
You're welcome!