Channel
Interviewed Person
The AI development landscape just got turned upside down. Vercel's new Agent Skills ecosystem is fundamentally changing how we work with Claude Code and other coding agents. The skills CLI tool let's you install skills with a single command, and the skills.sh marketplace let's you discover skills. But Cloudflare just dropped a competing RFC proposing (Agent Skills Discovery via Well-Known URIs) a completely different approach using .well-known URIs that could make Vercel's entire centralized directory model obsolete. š Relevant Links Cloudflare proposal - https://github.com/cloudflare/agent-skills-discovery-rfc?tab=readme-ov-file Vercel skills - https://vercel.com/changelog/introducing-skills-the-open-agent-skills-ecosystem Intro frontend design by shades 2134 - https://www.reddit.com/user/shades2134/ Remotion example from Magnus - https://x.com/mamagnus00/status/2014459283560358034 ā¤ļø More about us Radically better observability stack: https://betterstack.com/ Written tutorials: https://betterstack.com/community/ Example projects: https://github.com/BetterStackHQ š± Socials Twitter: https://twitter.com/betterstackhq Instagram: https://www.instagram.com/betterstackhq/ TikTok: https://www.tiktok.com/@betterstack LinkedIn: https://www.linkedin.com/company/betterstack š Chapters: 0:00 Intro 0:26 Agent Skills Recap 2:05 Vercel's add-skill project 2:58 Skills.sh for skill discovery 4:00 Cloudflare's Proposal 4:21 How Cloudflare's Proposal Works 5:11 Problems with the proposal
Various
Better Stack
Interviewed: v0
This is Skills by Bell, a directory of agent skills that you can add to any agent to extend its capabilities like making it amazing at React, turning it into a front-end designer, or even making it a motion graphics editor. The whole thing works by cloning and scanning a repo to see if it has a skills MD file before installing the skill. But will Cloudflare's new proposal make this amazing resource obsolete? Hit subscribe and let's get into it. Agent skills were introduced by Anthropic last year to improve Claude's
capabilities on specific tasks and works by adding a skills directory to the Claude one and then adding another directory with the name of the skill followed by a skills.md [music] file. Now, this file needs to contain a few key things for the skill to work. First, the top section which is front matter in YAML. That should contain at least the name of the skill and the description. Then below that outside of the front matter, this is where the instructions for the skill should go. The instructions get loaded into the model's
context so that it can be used for the specific task. And the beauty of skills lies in a technique called progressive disclosure where only the essential information is shown at first and the model can dig in to find even more if necessary. So when an agent loads up, only the information in the front matter of each skill is loaded into the context. And after it reads a prompt, it checks the description of each skill to find out if a skill can be used to enhance the prompt. And if it finds one, it then adds everything outside of the
front matter into the model's active context. You can even link other files to skills like references for a model to grab those if it thinks it's necessary. Since then, other companies like Microsoft, OpenAI, and Open Code have adopted agent skills, but there hasn't really been an easy way to install a skill for multiple agents without manually going into the relevant directory and cloning the repo. This became especially apparent when Versel wanted to share skills for their tools
for people to put in their specific agent. So, they made a project called add skill, which makes it easy to add a skill to 16 different agents. simply npx add skill followed by the name of the skill or the location of the repo. Now, I know what you're thinking. What's stopping people from using add skill to add any repo? Well, the way it works is after it passes the command, it then clones the repo, then looks for a skill.md file. If it can't find it, it cleans up by removing the repo. But if
it can find it, it auto detects a user's installed agents by looking through the configs. Then creates a sim link from agents/skills to wherever the skills need to be located for that agent. So doc cursor for cursor and dotclaws for claude code. It also has some telemetry for tracking which I'll talk about later. But at this point, Vel had a way for people to easily install skills to any agent. But what about discovering a skill? This is where skills.sh SH comes in showing currently a long list of
uncatategorized and unpadenated skills and the list gets populated I think by someone first installing a skill using ad skills or skills which kicks off the telemetry that's inside the package to add things like name agents and other information. The telemetry is completely anonymous and is not only used to populate the list, but also the amount of times a skill has been downloaded and by the specific agent, which can also be used to track the trending skills in the
last 24 hours. And also, if you're worried about being tracked, you can totally disable the telemetry. It's amazing how much work the cell labs have put into skill discovery and it's being used by loads of people to discover awesome skills like react native skills for your agents best practices for better or even the super popular remotion skill which people are using to create all kinds of crazy videos just from a single prompt. But has Cloudflare's new proposal made this all