This is a fascinating breakdown of Claude Skills, Simon! The idea that they boil down to a Markdown file with optional scripts is beautifully simple, and I totally see why you're excited about the potential 'Cambrian explosion.'
Your comparison to MCP really highlights the philosophical difference here. While I agree that Skills seem incredibly powerful and user-friendly for *individual* agent creation (especially with the low token overhead!), I wonder about the ecosystem angle compared to MCP. MCP's strength, as I see it, is that it provides a shared *developer* incentive—if I build an agent using the protocol, any LLM provider or user can immediately adopt it, fostering cross-platform compatibility.
Skills feel much more like a user-side empowerment tool, which is fantastic for personal productivity and customization.
I actually noticed that Claude was following some kind of documentation a while back!
There is really goes back to a best practice I haven't left ... Focus AI on a specific task or purpose that relies on specific, well-structured knowledge. Now you can just link those together.
I’ve been using Claude Code heavily and once you understand some core principles, it becomes a powerhouse!
Simplifying AI and AGI for everyone | This article breaks down complex ideas about Artificial Intelligence and Artificial General Intelligence into plain language, helping non-technical readers understand the key concepts as we move closer to the era of AGI.
Skills feels like a much better match for how LLMs work, than MCP. I never really understood the need for a new protocol for something which already talked fine to existing protocols (HTTP, CLIs, etc.).
The only practical issue with skills is that Claude Code struggles a bit in finding out when to use specific skills - unless you explicitly asks it to.
Claude Skills look like a real game-changer, simple, shareable, and powerful. The way they turn Markdown + scripts into specialized LLM abilities makes the agent concept feel much more tangible than MCP ever did.
For more AI trends and practical insights, check out my Substack where I break down the latest in AI.
CAN WRITERS COLLECT - $30 or $40K for all the many lawsuits?
Right now, as of 10/18/2025, there are 1.5 billion dollars out there waiting for writers who were hit by the AI corpro-rip-off of their works. The settlement means $3000.00 per each of their qualified books for the authors themselves, alongside their coauthors and publishers in equal splits.
WRITERS RIPPED OFF BY AI-CORPRO SUITIE CHCKEN SHITS
Can we writers quit raccooning long enough to get some hefty cash?
WATCH FOR SUDAY VESPERS ON OCTOBER 19, 2025 in order to continue to follow The Solid Footing Press News’s deep-dive into writers being ripped of by AI-suities. Sunday Vespers will take place every Sunday for the deep-dive. Put a posty on the frig!
NOTES FROM TESS—sizzling hot cub reporter for The Solid Footing Press News
Can writers quit racconing long enough to get a pile of cash dumped into their bank accounts?
Coming to this a bit late but your call on the Cambrian explosion has been playing out. Twenty-six tools have adopted the format since you wrote this. The fragmentation angle is what gets me; people are maintaining .cursorrules, .clinerules, CLAUDE.md, and skills folders side by side in the same project. Wrote up a proper overview of where the standard sits now and the mess it's trying to sort out: https://blog.devgenius.io/agent-skills-explained-teaching-your-ai-assistant-new-tricks-c6ee66a632c4
This is a fascinating breakdown of Claude Skills, Simon! The idea that they boil down to a Markdown file with optional scripts is beautifully simple, and I totally see why you're excited about the potential 'Cambrian explosion.'
Your comparison to MCP really highlights the philosophical difference here. While I agree that Skills seem incredibly powerful and user-friendly for *individual* agent creation (especially with the low token overhead!), I wonder about the ecosystem angle compared to MCP. MCP's strength, as I see it, is that it provides a shared *developer* incentive—if I build an agent using the protocol, any LLM provider or user can immediately adopt it, fostering cross-platform compatibility.
Skills feel much more like a user-side empowerment tool, which is fantastic for personal productivity and customization.
++ Good Post, Also, start here how to build tech, Crash Courses, 100+ Most Asked ML System Design Case Studies and LLM System Design
How to Build Tech
https://open.substack.com/pub/howtobuildtech/p/how-to-build-tech-10-how-to-actually?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/howtobuildtech/p/how-to-build-tech-06-how-to-actually?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/howtobuildtech/p/how-to-build-tech-05-how-to-actually?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/howtobuildtech/p/how-to-build-tech-04-how-to-actually?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/howtobuildtech/p/how-to-build-tech-03-how-to-actually?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/howtobuildtech/p/how-to-build-tech-01-the-heart-of?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/howtobuildtech/p/how-to-build-tech-02-how-to-actually?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
Crash Courses
https://open.substack.com/pub/crashcourses/p/crash-course-04-hands-on-crash-course?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/crashcourses/p/crash-course-03-hands-on-crash-course?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/crashcourses/p/crash-course-02-a-complete-crash?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/crashcourses/p/crash-course-01-a-complete-crash?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
LLM System Design
https://open.substack.com/pub/naina0405/p/very-important-llm-system-design-499?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/naina0405/p/very-important-llm-system-design-63c?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/naina0405/p/very-important-llm-system-design-bdd?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/naina0405/p/very-important-llm-system-design-661?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/naina0405/p/very-important-llm-system-design-83b?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/naina0405/p/very-important-llm-system-design-799?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/naina0405/p/very-important-llm-system-design-612?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/naina0405/p/very-important-llm-system-design-7e6?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/naina0405/p/very-important-llm-system-design-67d?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/naina0405/p/most-important-llm-system-design-b31?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://naina0405.substack.com/p/launching-llm-system-design-large?r=14q3sp
https://naina0405.substack.com/p/launching-llm-system-design-2-large?r=14q3sp
[https://open.substack.com/pub/naina0405/p/llm-system-design-3-large-language?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/naina0405/p/important-llm-system-design-4-heart?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/naina0405/p/very-important-llm-system-design-63c?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
They are maybe my favorite tool right now in llms
I actually noticed that Claude was following some kind of documentation a while back!
There is really goes back to a best practice I haven't left ... Focus AI on a specific task or purpose that relies on specific, well-structured knowledge. Now you can just link those together.
I’ve been using Claude Code heavily and once you understand some core principles, it becomes a powerhouse!
Simplifying AI and AGI for everyone | This article breaks down complex ideas about Artificial Intelligence and Artificial General Intelligence into plain language, helping non-technical readers understand the key concepts as we move closer to the era of AGI.
https://open.substack.com/pub/cyberamyntas/p/the-crown-jewels-of-agi?r=404gdk&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
Skills feels like a much better match for how LLMs work, than MCP. I never really understood the need for a new protocol for something which already talked fine to existing protocols (HTTP, CLIs, etc.).
The only practical issue with skills is that Claude Code struggles a bit in finding out when to use specific skills - unless you explicitly asks it to.
++ Good Post, Also, start here 100+ Most Asked ML System Design Case Studies and LLM System Design
https://open.substack.com/pub/naina0405/p/bookmark-most-asked-ml-system-design?r=14q3sp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
Claude Skills look like a real game-changer, simple, shareable, and powerful. The way they turn Markdown + scripts into specialized LLM abilities makes the agent concept feel much more tangible than MCP ever did.
For more AI trends and practical insights, check out my Substack where I break down the latest in AI.
There’s this too with all those Awesome Skills:
$1.5 BILLION JUST WAITING FOR WRITERS
RIPPED OFF BY AI-CORPRO
HOW MANY CLASS ACTION LAWSUITES ARE THERE?
CAN WRITERS COLLECT - $30 or $40K for all the many lawsuits?
Right now, as of 10/18/2025, there are 1.5 billion dollars out there waiting for writers who were hit by the AI corpro-rip-off of their works. The settlement means $3000.00 per each of their qualified books for the authors themselves, alongside their coauthors and publishers in equal splits.
WRITERS RIPPED OFF BY AI-CORPRO SUITIE CHCKEN SHITS
Can we writers quit raccooning long enough to get some hefty cash?
WATCH FOR SUDAY VESPERS ON OCTOBER 19, 2025 in order to continue to follow The Solid Footing Press News’s deep-dive into writers being ripped of by AI-suities. Sunday Vespers will take place every Sunday for the deep-dive. Put a posty on the frig!
NOTES FROM TESS—sizzling hot cub reporter for The Solid Footing Press News
Can writers quit racconing long enough to get a pile of cash dumped into their bank accounts?
Skills makes so much sense
I think they are complementary. MCP is deterministic. Skills seem to be non deterministic The context bloat created by MCP is fixable.
Coming to this a bit late but your call on the Cambrian explosion has been playing out. Twenty-six tools have adopted the format since you wrote this. The fragmentation angle is what gets me; people are maintaining .cursorrules, .clinerules, CLAUDE.md, and skills folders side by side in the same project. Wrote up a proper overview of where the standard sits now and the mess it's trying to sort out: https://blog.devgenius.io/agent-skills-explained-teaching-your-ai-assistant-new-tricks-c6ee66a632c4
Great breakdown