Subscribe to our newsletter
One short email when we publish. No spam, unsubscribe anytime.

The 15 most-starred skills for Hermes Agent, checked one by one
On September 30, 2026 the X account HermesWatcher posted a list it presented as the 15 most-starred skills for Hermes Agent. By the morning of October 1 it had 2,708 bookmarks. A skill is a folder of plain-text instructions an AI agent reads when a task needs it. The reasons below are star counts as of October 1, 2026 and what each maker's own page says.
The 15, in star order
1. Superpowers
What it is: Superpowers is a set of skills plus a method for building software with an agent. The agent asks what the user wants, works in small steps and writes a test before the code.
Why it is popular: It holds 293,554 stars, the most of the 15, a star being one GitHub user's public bookmark. It has been on GitHub since October 2025. Its README, the note a project shows first, carries install steps for Hermes Agent.
2. Anthropic's skills collection
What it is: Anthropic's skills collection is the public skill set of Anthropic, maker of the Claude models. Its 19 folders include skills for Word, PDF, PowerPoint and Excel files.
Why it is popular: Second with 179,214 stars, it comes from Anthropic itself. Hermes Agent subscribes to it by default, per the Hermes Agent documentation.
3. UI UX Pro Max
What it is: UI UX Pro Max gives an agent design guidance for app and web screens. It covers style, colours, fonts and layouts.
Why it is popular: Third with 132,096 stars, it has been on GitHub since November 2025. Its README lists 79 design styles and 22 tech stacks, the toolkits apps are built with.
4. Graphify
What it is: Graphify, from Graphify Labs, maps how a project's code files and documents connect. The agent can then ask the map questions.
Why it is popular: It gathered 122,862 stars in about six months, on GitHub since April 2026. Only number 1 gained stars faster per day, by OneClickClaw's count.
5. claude-mem
What it is: claude-mem gives the agent a memory from one conversation to the next.
Why it is popular: The oldest of the 15, on GitHub since August 2025, it holds 95,040 stars. Its description is the only one naming both OpenClaw and Hermes Agent.
6. Archify
What it is: Archify draws architecture, workflow and data-flow diagrams, each as a single web page file that opens in a browser.
Why it is popular: It reached 75,401 stars in under six months, on GitHub since April 2026. Its README says the file opens with nothing else installed and can be sent to someone else.
7. last30days
What it is: last30days researches a topic across Reddit, X, YouTube, Hacker News and the web, then writes a summary.
Why it is popular: It holds 63,293 stars and has been on GitHub since January 2026. Its description says the summary is grounded, meaning tied to the sources it found.
8. OpenMontage
What it is: OpenMontage is a video production system run by an agent.
Why it is popular: It collected 62,049 stars in about six months, on GitHub since March 2026. Its description counts 12 production pipelines, or video workflows, and more than 100 tools.
9. Humanizer
What it is: Humanizer removes the tell-tale signs of machine-written text, the phrasing that gives a program away.
Why it is popular: Since January 2026 it has drawn 53,193 stars on GitHub. It ships as a single skill file.
10. I Have ADHD
What it is: I Have ADHD makes the agent put the answer first, number the steps and cut the closing filler.
Why it is popular: The youngest of the 15, since May 2026, it picked up 52,488 stars. Its README, in ten languages, says it stops the agent burying the answer.
11. Marketing Skills
What it is: Marketing Skills covers copywriting, search, analytics, paid ads and page conversion, meaning turning visitors into customers.
Why it is popular: On GitHub since January 2026, it stands at 52,049 stars and holds 50 skill files.
12. Planning with Files
What it is: Planning with Files keeps the plan of a long task in files on disk. The agent can then pick it up after an interruption.
Why it is popular: It holds 27,216 stars and has been on GitHub since January 2026. Its description says it works with more than 60 agents.
13. SkillSpector
What it is: SkillSpector is a security scanner that checks a skill for risky patterns before it is installed.
Why it is popular: It comes from NVIDIA, the chip company, and holds 18,833 stars. Its description lists hidden instructions and data leaks among the risks it looks for.
14. Skill Seekers
What it is: Skill Seekers converts documentation sites, code projects and PDF files into skills.
Why it is popular: It counts 15,069 stars and has been on GitHub since October 2025.
15. BrowserAct skills
What it is: BrowserAct skills let an agent drive a web browser, opening websites, reading pages and pulling out data. They come from BrowserAct, which makes browser tools for agents.
Why it is popular: On GitHub since February 2026, the collection holds 6,062 stars, the fewest on the list. It contains 103 skill files.
What the check found
OneClickClaw read the public GitHub record of all 15 on October 1, 2026. All 15 exist and are public, and the post's order matches the real star order.
Together the 15 hold about 1.25 million stars, the first about 48 times the fifteenth. Fourteen contain a file named SKILL.md, which tells the agent what the skill is and when to use it. Number 4 names that file in lowercase.
Ten use the MIT licence and three Apache 2.0, both allowing almost any use with credit. Number 8 uses AGPL 3.0, which asks anyone who changes the software and serves it online to publish those changes. Number 2 has no single licence for the whole project.
Before installing one
Most makers' descriptions name other agents too, so these are general skills in one shared format, the Agent Skills specification. Both official docs were read on October 1, 2026.
The Hermes Agent documentation says the agent can install a skill straight from a public GitHub project. Every skill installed through its hub is scanned for data leaving the machine, hidden instructions and destructive commands. Findings marked dangerous stay blocked.
The OpenClaw documentation says it follows the same standard. Skills install from its catalogue, ClawHub, or a project's code folder. It tells users to treat third-party skills as untrusted code and read them before switching them on.
A skill works only while the agent runs
Per the Hermes Agent documentation, the agent loads a skill only when a task needs it. So a skill helps only while the agent runs. OneClickClaw hosts Hermes Agent and OpenClaw on a private server that stays on. The Starter plan is EUR 14.99 per month, with a seven-day free trial, at oneclickclaw.io/hermes.
Source: x.com
Helpful documentation
Want your own Hermes agent running without the setup? We host it on a dedicated EU server. Free for 7 days, no credit card.
Start your free trialGet notified when we publish new articles
No spam, unsubscribe anytime.
By subscribing, you agree to our Privacy Policy.
