Guide · Agent skill
How to use the Era Alpha stock research agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
Era Alpha is an agent skill, a packaged set of written instructions that an AI agent such as Claude or Codex loads so it follows a specific research method. This guide explains what the era-alpha skill makes the agent do when you ask about a sector, how to install it, how to phrase requests, and what it needs. By the end you will be able to run a structured, written research process on a fast-growing industry and review its output critically.
Era Alpha Stock Researchxbtlin/ai-berkshire · MIT Download the skill Quick start
- 1Download the ZIP. Download the era-alpha ZIP from the skill page on this site and unzip it.
- 2Place the folder. Move the era-alpha folder into ~/.claude/skills/ for personal use, or into .claude/skills/ inside a project.
- 3Or upload in the Claude apps. In the Claude apps, open Settings, go to Capabilities, then Skills, and upload the skill.
- 4Prepare the repository. Make sure you have a local checkout of the ai-berkshire repository so the agent can read AGENTS.md and run the tools/ scripts.
- 5Make a natural request. Ask about a specific sector or industry and say whether you want a full report or a narrower check.
- 6Review the report. Check the data cutoff date, verify key figures against primary sources, and set up the listed warning signals for monitoring.
What the era-alpha skill makes the agent do
The skill is based on a four-step retail-investing method: build a map of the industry, ask the core questions, verify the thesis across several data sources, then decide on valuation and holding rules. Its goal is to identify leading companies in a high-growth sector that appear to have pricing power, a moat and sustainable growth, and to set explicit conditions for selling.
When you ask about a sector, the agent narrows the field to two or three core segments of the supply chain instead of covering every company. For the leading companies in those segments, it cross-checks quarterly financial reports against high-frequency indicators such as shipments, prices and orders, assesses whether valuation is reasonable, and produces a report with observable warning signals.
- Supply-chain map with the stage, competitive landscape and candidate leaders for each segment
- Core-question answers that rely on the agent's own data rather than market consensus
- Growth-sustainability checks across financial reports, high-frequency data, industry tracking, competition and macro conditions
- Valuation view and entry approach, plus a list of thesis-breaking signals
- A self-critique section listing at least three ways the report could be wrong
Install the era-alpha skill from the ZIP download
Download the ZIP from the skill page on this site, then unzip it. Put the era-alpha folder in your personal skills directory at ~/.claude/skills/, or in a project-level .claude/skills/ folder inside the project you work in. Claude Code and Codex users share the same canonical workflow, so the skill file is written to work in either environment.
If you use the Claude apps instead, open Settings, go to Capabilities, then Skills, and upload the skill there. The skill's only file is SKILL.md, which contains the full instructions.
- Personal install: ~/.claude/skills/
- Project install: .claude/skills/ inside your project folder
- Claude apps: Settings → Capabilities → Skills, then upload
~/.claude/skills/Set up the repository the skill depends on
The skill refers to the ai-berkshire repository. It tells the agent to read AGENTS.md for research quality rules, to use shared scripts in the repository's tools/ folder, and to follow the financial-data skill for cross-checking. Without access to that repository, the agent cannot run those tools or follow those rules, so plan to have a local checkout of the repository available.
The skill's instructions say to run commands from the repository root, using paths such as the one below. If your session starts outside the repository, the agent is told to locate the actual checkout first rather than assume a fixed path.
- Run the date command first to confirm today's date, and state the data cutoff in the report header
- Use exact arithmetic tools for valuation and maths
- Label uncertainty and source gaps clearly in the report
python3 tools/financial_rigor.pyTrigger the skill with a natural request
You do not need special syntax. Name the sector or industry you want researched and say what you want, such as a sector overview, a comparison of leading companies, or an assessment of whether a segment is still growing. The skill is written to take your request as the sector or direction to analyse.
Because the skill is in Chinese, you can ask in English and receive a report in English, but the source method and its section names are in Chinese. If you want the report in Chinese, say so in your request.
- Analyse the AI server supply chain and tell me which two or three segments matter most
- Who are the leaders in the solid-state battery segment, and do they have pricing power?
- Using the era-alpha approach, research the leading companies in the humanoid robot sector and list the warning signals for selling
- Build a supply-chain map for the photovoltaic inverter industry and give me a valuation view
What a finished report contains
The report follows a fixed structure, so you can compare runs. Each section is named by analytical dimension rather than by any investor's name. Expect dense data tables and explicit statements where signals conflict.
Treat every number as something to check. The skill itself says that financial reports are a lagging picture and that high-frequency data should be used to correct them, but the agent still depends on web search results, which can be incomplete or out of date.
- 1. Industry map: segment table and the reason for selecting core segments
- 2. Answers to the core questions and the reasoning behind the chosen leader
- 3. Growth-sustainability checks with contradictory signals stated openly
- 4. Valuation anchor and entry recommendation
- 5. Holding discipline and an observable, falsifiable list of thesis-breaking signals
- 6. Ways this report could be wrong
Safety notes: scripts and network access
This skill runs scripts and uses the network. It calls shell commands to run tools in the repository, such as financial_rigor.py, and it uses web search to fetch current financial reports and data. Review the repository's tools/ scripts before you run the skill, and only use it in an environment where you are comfortable with those commands running.
The output is analysis, not financial advice. The skill is designed for individual investors and analysts, and it does not come with a verified track record. Check every key figure against the original filing before acting on it.
- Read tools/ scripts before first use
- Verify growth rates, gross margins and valuation percentiles against primary sources
- Note the data cutoff date in each report
Tips for better results
Narrow the scope. The method works best on one industry with a small number of segments. Asking for a whole market produces a shallow report, because the skill is designed to go deep on two or three segments.
Expect a slow answer. The workflow is multi-step and can need many searches and parallel sub-agents, so it is not suited to a quick answer. Use the thesis-breaking signals as a monitoring checklist and revisit them on the cadence the report recommends.
- Name one sector and, if you have one, the companies you already follow
- Ask for the report in Chinese if you need it in that language
- Use the exit-signal list as a recurring review, not a one-off task
What people use it for
- Mapping the supply chain of a fast-growing sector and choosing two or three segments to study in depth
- Comparing leading companies within one industry on pricing power, moat and growth quality
- Building a written entry and exit plan with explicit, observable warning signals
- Producing a Chinese-language sector report for a Chinese-speaking reader
Frequently asked questions
Is the era-alpha skill free?
The skill is published under the MIT licence, which is a permissive open-source licence. The source does not state any price, so check the repository licence file for the terms that apply to you.
What do I need to run the era-alpha skill?
You need an agent environment that supports skills, web search access for current financial reports, and shell access so the agent can run the repository's tools/ scripts such as financial_rigor.py. You also need access to the ai-berkshire repository, because the skill refers to AGENTS.md and the financial-data skill.
Does the era-alpha skill give trading advice?
No. The source describes the output as analysis and not financial advice. It relies on web research that can be incomplete or outdated, so you should verify the data yourself before making any trade.
How do I install era-alpha in Claude?
Download the ZIP from the skill page, unzip it, and place the folder in ~/.claude/skills/ for personal use or in .claude/skills/ inside a project. In the Claude apps, you can instead open Settings, go to Capabilities, then Skills, and upload it.
Why is the skill slow or using many searches?
The workflow is multi-step by design. It can require many web searches and parallel sub-agents to gather financial reports, high-frequency data and industry information. For a quick answer, narrow your request to a single segment.
How does era-alpha compare with other research skills?
The skill is narrower and deeper than the industry-research skill, which focuses on a full supply-chain overview, and the industry-funnel skill, which screens the whole market. Era-alpha concentrates on identifying leading companies in a high-growth segment and setting explicit exit signals.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use era-alpha if you want a structured, written research process for one fast-growing sector, need explicit thesis-breaking signals, or want the report in Chinese. Pick something else if you need a quick answer, research beyond listed and private companies and their valuation, or a verified track record, because the source provides none.
Download the skill