COOLDOWN PERIOD

[hesreallyhim]

Cooldown period in effect to mitigate influx of recommendations of the following sort:

• authored and submitted by coding agents
• projects that have 0 stars and no commits older than 24 hours (possibly this is due to prototyping in a private repo and then releasing in a new one - I do this too, that's fine... but I wouldn't submit it to an Awesome list on day 1)
• projects having to do with DeFi and crypto - mentioned in pinned Issue I cannot accept these at the moment, will make that clearer
• loads of people trying desperately to submit a recommendation via the gh CLI - due to limitations of the CLI and the architecture of this repo, this is technically not feasible
• submissions coming in through PRs - more evidence that people are not reading the CONTRIBUTING docs
• recommendations without sufficient validation instructions
• "Flavor of the week" projects
• Projects that are tangentially related to Claude Code (Claude-Code-compatible)
• Specifically: Loads of projects that do not focus on Claude Code functionality, but instead on how to hook up Claude Code to something else - no judgment against these projects but not a priority for a list about awesome things to do with Claude Code
• Projects that require API keys
• Orchestrators as far as the eye can see
• Projects at version 0.0.0.0.1

I have seen some very well designed projects submitted for recommendation. These authors are getting drowned out by the noise. I will spend some time "grooming" the recommendation list. Probably some very good projects will be removed unfairly. That's really disappointing to me. But the current situation is that these projects are sitting around without review anyway.

I will also spend some time going through existing projects on the list and removing entries that are no longer maintained or relevant.

Please have respect for the authors on this list. It's really great that people want to share their work here, and I want to make sure that people continue to do so. Submitting projects that don't meet the standards of existing entries is detrimental to your fellow developers. Although sometimes I get lucky and find a hidden gem - I should not be the first person to try your library.

Take a look at what Anthropic is releasing - Claude Code is amazing. The purpose of this list is to highlight those features. There are new features all the time - projects that explore these features are ones that are in the spirit of this repo.

This is a hard thing to face but it can't be denied: Claude Code can produce plugins that are as good as many of the entries on this list (undoubtedly, it's fair to say that many of the submissions probably are already authored by Claude Code). Once you have a powerful skill-creator skill, or a plugin-creator plugin, this fundamentally changes the landscape of what is possible, and also of what is noteworthy. It raises the bar significantly for what is an awesome project. It's hard for me to recommend something that I know Claude can build for me just as easily. When skills were released, a really well-written skill was something very valuable. But the functionality has matured - what skills are there that can't be produced through a few iterations of prompting and improving? (That's not a rhetorical question - that's a challenge.)

Perhaps the only way to stand out, then, is to be creative, and to build something that it would not occur to me to ask Claude to build. Or to actually build something that has a specific use value and is not just Claude-Code-for-Claude-Coders.

I don't know the best way to approach these problems. Surely, for new users, seeing really good slash commands must be something of value. But nobody is submitting just slash commands. The Claude Code documentation site is primarily about plugins, agents, and skills. I really am curious to know how someone new to Claude Code is supposed to understand what these things do if they don't have a deep understanding, and good examples, of the basic components. (Or maybe they don't need to, so long as they have a command-creator skill.)

I am considering additional ways to organize the list to accommodate the changes in developer tendencies that have emerged, and the overall shift in Claude-Code-engineering that has occurred.

Maybe the problem I'm facing is a little bit like what is described in this post by Anthropic regarding designing a suitable take-home problem for hiring engineers: "We were about to release a model where the best strategy on our take-home would be delegating to Claude Code."

I'm very confident that there are a lot of interesting things to build that can't just be delegated to Claude - or, rather, if they can, the hard part is primarily no longer in "building things" but in making them interesting. So that's what I would encourage people to do if they want recognition on this list. After all, there wasn't anything ground-breaking about "Wiggum" - but it was interesting and clever. But building the Enhanced Battle-Tested Ultimate Ralph Wiggum Super-Plugin... you see where I'm going, I hope.