all-mpnet-base-v2
sentence-transformers · Kit
Turns sentences and paragraphs into 768-number vectors so you can search, cluster and compare text by meaning. It suits developers and data teams working in En…
Guides an AI agent through an industry-chain method for stock and fund research: it splits a hot theme into layers, finds supply bottlenecks, checks filings and risks, and ranks what to study first.
Serenity.skill is an open-source project on GitHub in the Finance & accounting category, listed by Mindket. It is free. This is version 1, last updated on . Install it from GitHub as its README says. Get Mindket’s setup notes on your Mindket account, as a zip.
SERENITY_DIR="$HOME/.claude/skills/serenity-skill" mkdir -p "$SERENITY_DIR" cp -R SKILL.md LICENSE references assets examples agents "$SERENITY_DIR"/
Serenity.skill is an agent skill for stock and fund research. It starts from a hot theme such as AI semiconductors or robotics and splits the industry into layers, looking for the steps hardest to expand or replace. Candidate companies and funds are then checked against filings, earnings and risks, and you get a ranked list of what to research first.
It suits investors who watch a constant stream of hot themes and want a repeatable screening routine. The AI does the first round of research and you decide what to dig into.
| Industry | Who uses it | How they use it | Example |
|---|---|---|---|
| Individual investing | Retail investors | Turn a trending theme into a short list of areas and companies to study first | Ask which AI semiconductor layers sit closest to real demand |
| Fund and ETF selection | Theme fund buyers | Find the chain behind a fund and which top holdings to verify | Check whether a robotics fund holds companies near the bottleneck |
| Equity research | Analysts and students | Challenge a claim about a company's place in a supply chain | Test whether a company is truly a core CPO supplier |
| Investor learning | Newcomers to chain research | Be walked through the method one question at a time | Ask to be taught the method, one question per turn |
If you want live quotes, price alerts or a buy signal, skip it. It brings no market data and leaves the final decision to you.
| Use case | Who | What they get back |
|---|---|---|
| Theme screening | Investors facing a hot topic | The industry split into layers, with the ones nearest real demand and tight capacity ranked first |
| Company challenge | Someone holding a stock tip | The company's real chain position, customer evidence, revenue quality and main risks |
| Fund review | Buyers of theme funds | The core chain behind the fund and the holdings worth verifying |
| Candidate ranking | Investors with a few stocks | A research order based on chain position, evidence strength, valuation pressure and risk |
For Claude Code, copy the files into your user skills folder:
SERENITY_DIR="$HOME/.claude/skills/serenity-skill" mkdir -p "$SERENITY_DIR" cp -R SKILL.md LICENSE references assets examples agents "$SERENITY_DIR"/
For Codex, the same copy goes to a different folder:
SERENITY_DIR="$HOME/.agents/skills/serenity-skill" mkdir -p "$SERENITY_DIR" cp -R SKILL.md LICENSE references assets examples agents "$SERENITY_DIR"/
To use it in one project only, point SERENITY_DIR at the project's .claude/skills/serenity-skill folder for Claude Code, or .agents/skills/serenity-skill for Codex.
Call the skill by name and say what to research. Claude Code starts the call with a slash and Codex with a dollar sign.
/serenity-skill 挑战天孚通信是 CPO 核心供应商的说法,逐项核对原始披露。 $serenity-skill 研究现在 A 股 AI 半导体产业链,说明优先研究方向、证据和反方理由。
The README's sample prompts are in Chinese, and an English README sits beside it.
| Tool | What it does |
|---|---|
| /serenity-skill | Calls the skill in Claude Code |
| $serenity-skill | Calls the skill in Codex |
| research-prompt-pack.md | More copy-ready prompt templates |
| thesis-template.md | A template for a full research memo: chain position, confirmed facts, missing evidence, profit and valuation, substitutes, failure conditions |
| validate_skill.py | An optional check of the skill's name, description and folders, not of any investment view |
"Use serenity-skill to research robotics, split the chain first, then list the links closest to a real supply bottleneck," you could say. Its output format opens with the directions to look at first, then lists companies, each with a chain position and something to track such as orders or margins. It also marks down a popular direction that still lacks proof like orders or profit, and ends with three checks for next.
Think of it as a routine your AI follows when a hot topic shows up. Rather than asking which stock is hot, it first asks where real demand sits and which step is hardest to scale. A good first try is to install it and ask it to challenge a company someone recommended to you.
It is a method written as skill files: SKILL.md plus reference notes on the research workflow, an evidence ladder, market sources, and risk and compliance. It screens for few suppliers, long qualification cycles, hard capacity expansion, strict customer certification and demanding material purity. Priorities rest on stated evidence and reasoning, with no combined numeric score. The tradeoff is that it brings no data feed, so results depend on what your client can search and read.
# Serenity.skill Guides an AI agent through an industry-chain method for stock and fund research: it splits a hot theme into layers, finds supply bottlenecks, checks filings and risks, and ranks what to study first. Made by muxuuu on GitHub. These notes are Mindket's, written from its README; the project's own files and license are in its repository. ## About What it is: Serenity.skill is an agent skill for stock and fund research. It starts from a hot theme such as AI semiconductors or robotics and splits the industry into layers, looking for the steps hardest to expand or replace. Candidate companies and funds are then checked against filings, earnings and risks, and you get a ranked list of what to research first. Who it suits: It suits investors who watch a constant stream of hot themes and want a repeatable screening routine. The AI does the first round of research and you decide what to dig into. ```table Industry | Who uses it | How they use it | Example Individual investing | Retail investors | Turn a trending theme into a short list of areas and companies to study first | Ask which AI semiconductor layers sit closest to real demand Fund and ETF selection | Theme fund buyers | Find the chain behind a fund and which top holdings to verify | Check whether a robotics fund holds companies near the bottleneck Equity research | Analysts and students | Challenge a claim about a company's place in a supply chain | Test whether a company is truly a core CPO supplier Investor learning | Newcomers to chain research | Be walked through the method one question at a time | Ask to be taught the method, one question per turn ``` If you want live quotes, price alerts or a buy signal, skip it. It brings no market data and leaves the final decision to you. Where to use it: - Claude Code: install it in your user skills folder or a project's skills folder, then call it with a slash command. - Codex: the same idea, called with a dollar sign command. - Other clients that read skill files: they use the same skill, and the install location follows each client's own documentation. - Your client's own web search, browser or filings tools supply the data the research needs. ```table Use case | Who | What they get back Theme screening | Investors facing a hot topic | The industry split into layers, with the ones nearest real demand and tight capacity ranked first Company challenge | Someone holding a stock tip | The company's real chain position, customer evidence, revenue quality and main risks Fund review | Buyers of theme funds | The core chain behind the fund and the holdings worth verifying Candidate ranking | Investors with a few stocks | A research order based on chain position, evidence strength, valuation pressure and risk ``` How to set it up: - An agent client that reads skill files, such as Codex or Claude Code - The client's own web search, browser or filings data tools, because the skill brings no market data - Python 3, only for the optional structure check 1. Get the repository by cloning it or by downloading the ZIP and unpacking it. 2. Open a terminal in the folder that holds SKILL.md. 3. Run the copy commands below for your client. 4. If an older copy exists, move it out of the skills search folders first. Copying over it keeps retired files, and two copies may load twice. 5. Open a new session and check that the client finds serenity-skill. For Claude Code, copy the files into your user skills folder: ```bash SERENITY_DIR="$HOME/.claude/skills/serenity-skill" mkdir -p "$SERENITY_DIR" cp -R SKILL.md LICENSE references assets examples agents "$SERENITY_DIR"/ ``` For Codex, the same copy goes to a different folder: ```bash SERENITY_DIR="$HOME/.agents/skills/serenity-skill" mkdir -p "$SERENITY_DIR" cp -R SKILL.md LICENSE references assets examples agents "$SERENITY_DIR"/ ``` To use it in one project only, point SERENITY_DIR at the project's .claude/skills/serenity-skill folder for Claude Code, or .agents/skills/serenity-skill for Codex. How to use it: Call the skill by name and say what to research. Claude Code starts the call with a slash and Codex with a dollar sign. ```text /serenity-skill 挑战天孚通信是 CPO 核心供应商的说法,逐项核对原始披露。 $serenity-skill 研究现在 A 股 AI 半导体产业链,说明优先研究方向、证据和反方理由。 ``` The README's sample prompts are in Chinese, and an English README sits beside it. ```table Tool | What it does /serenity-skill | Calls the skill in Claude Code $serenity-skill | Calls the skill in Codex research-prompt-pack.md | More copy-ready prompt templates thesis-template.md | A template for a full research memo: chain position, confirmed facts, missing evidence, profit and valuation, substitutes, failure conditions validate_skill.py | An optional check of the skill's name, description and folders, not of any investment view ``` Example: "Use serenity-skill to research robotics, split the chain first, then list the links closest to a real supply bottleneck," you could say. Its output format opens with the directions to look at first, then lists companies, each with a chain position and something to track such as orders or margins. It also marks down a popular direction that still lacks proof like orders or profit, and ends with three checks for next. For newcomers and experts: Think of it as a routine your AI follows when a hot topic shows up. Rather than asking which stock is hot, it first asks where real demand sits and which step is hardest to scale. A good first try is to install it and ask it to challenge a company someone recommended to you. It is a method written as skill files: SKILL.md plus reference notes on the research workflow, an evidence ladder, market sources, and risk and compliance. It screens for few suppliers, long qualification cycles, hard capacity expansion, strict customer certification and demanding material purity. Priorities rest on stated evidence and reasoning, with no combined numeric score. The tradeoff is that it brings no data feed, so results depend on what your client can search and read. ```flow Pick a hot theme Split the chain into layers Find the hardest-to-expand steps Screen companies and funds Check filings, customers, capacity, risk Get a ranked research list ``` Good to know: - Research support only: it ranks and reasons, and the buy or sell decision stays with you. - It has no live quotes or data account, so your client needs its own search, browser or filings tools. - Company judgments are meant to trace back to filings, exchange documents, earnings reports and calls, regulatory files, patents, standards, trusted media and professional analysis. - The two full examples are real public-source studies as of 2026-09-14, with their sources and unfinished checks. They show the method and are not a live list or a buy list. - Testing so far: Codex CLI 0.147.0 loaded this version and ran a company-claim challenge on offline material. Claude Code's folder layout was checked, but model calls were not tried. - The method is drawn from research paths visible in Serenity's public posts. - License: MIT. ## On GitHub, 16 November 2026 - Stars: 4,170 - Forks: 636 - Watching: 15 - Open issues and pull requests: 3 - Language: Python - License: MIT - Topics: agent-skills, ai-agents, claude-code, codex, investment-research, stock-research, supply-chain - Trending: #22 repository of the day on GitHub trending (16 November 2026) ## Links - Code: https://github.com/muxuuu/serenity-skill - Releases: https://github.com/muxuuu/serenity-skill/releases - Issues: https://github.com/muxuuu/serenity-skill/issues - Star history: https://www.star-history.com/#muxuuu/serenity-skill&Date ## Get the code ``` git clone https://github.com/muxuuu/serenity-skill.git ``` Then follow the setup in its README. Nothing in this kit runs by itself.
README.md 8 KB
sentence-transformers · Kit
Turns sentences and paragraphs into 768-number vectors so you can search, cluster and compare text by meaning. It suits developers and data teams working in En…
intfloat · Kit
Turns text in about 100 languages into vectors so you can compare meaning across them, for search and matching. Made for developers who build multilingual retr…
aldegad · Skill
Turns one base drawing of a character into game-ready sprites, either an atlas with a manifest or transparent motion loops, for game developers working in Code…
Leonxlnx · Skill
Leonxlnx’s Agent Skill that gives Claude Code and Codex an art direction, pictures from open collections and a real-browser check, so the websites and app scre…
tt-a1i · Skill
Generates interactive diagrams and explanatory pages from codebases or textual plans. It allows AI agents to visualize workflows, architecture, and complex log…
coreyhaines31 · Skill
A set of marketing skills for coding agents, covering conversion, copywriting, SEO, analytics and growth. Built for technical marketers and founders who use Cl…