Is Buffett Skills Repository Agi Worth It In 2026? A Complete Review

You are staring at a 200-page 10-K filing for a boring industrial company, trying to figure out if it has a genuine moat or if management is just blowing smoke about "synergies." Usually, you’d dump the text into a standard AI and get a generic summary that sounds like a high school business report. It misses the nuance of owner earnings, ignores the institutional imperative, and fails to calculate a real margin of safety.

That is the specific headache buffett skills Repository agi now buffett skills aims to cure. By injecting Warren Buffett’s entire mental model directly into your terminal via Claude Code, it transforms a standard LLM into a cynical, value-obsessed analyst. I spent the last week running my portfolio through this repository to see if it actually uncovers hidden gems or if it is just a fancy prompt wrapper that costs too much to run.

What is buffett skills Repository agi now buffett skills?

buffett skills Repository agi now buffett skills is an AI developer tool and specialized skill set for Claude Code that implements Warren Buffett's investment framework to enable deep financial analysis and stock evaluation through structured prompts and reference data—distinguishing itself by using a progressive disclosure architecture that loads specific financial mental models only when needed.

Built by the agi-now team, this isn't a standalone app or a web-based stock screener. It is a set of instructions and reference files designed for the Claude Code CLI. It solves the "hallucination and shallowness" problem by forcing the AI to consult 49 distinct concept pages derived from decades of Buffett’s shareholder letters before it gives you an opinion. It effectively moves the AI from "guessing what Buffett would say" to "checking what Buffett actually wrote."

Hands-on Experience: Testing the Buffett Framework

The Workflow: From Terminal to Verdict

Using this tool feels significantly different from chatting with a standard AI. Because it lives inside Claude Code, the "buffett" skill is invisible until you trigger it. I tested this by opening a directory containing a PDF of a recent annual report and simply typing claude "Should I buy this stock?".

The first thing you notice is the "tool call" activity. Unlike a standard prompt that replies in seconds, this skill is methodical. It triggers a series of background reads, pulling from its internal reference files on "Moats," "Management Assessment," and "Owner Earnings." You can see it working through the logic: it doesn't just look at the P/E ratio; it searches the text for signs of capital allocation skill. When I tested it on a complex banking stock, it correctly identified the risks in the loan book that a generic model completely ignored.

Analytical Depth vs. Standard AI

The standout feature is the progressive disclosure architecture. Most AI tools try to cram every instruction into one massive prompt, which leads to the model forgetting half of them. This repository keeps the main SKILL.md lean and only fetches the "Banking Framework" or "Insurance Playbook" when it detects you are looking at those specific industries.

In my tests, the output was consistently structured. You don't get a rambling essay. You get a "Financial Snapshot" with ROIC and cash conversion rates, followed by a "Margin of Safety" calculation. The most impressive part is the "Sell Criteria Check." Most investors (and AI tools) focus on when to buy; this tool forces the AI to evaluate the four specific conditions Buffett uses to exit a position, which is a level of discipline I haven't seen in other AI financial agents.

Where the Polish Wears Thin

It is not all smooth sailing. Because the tool is so thorough—making 11+ tool calls compared to the usual 2—it is slow. You will wait nearly three minutes for a full analysis. Furthermore, while the "Buffett Voice" in the final verdict is fun, it can sometimes feel a bit like a caricature. I found that if the data in the financial report is messy or non-standard, the tool occasionally struggles to map "owner earnings" correctly, requiring you to manually point it to the right tables in the document.

Getting Started with the Repository

To use buffett skills Repository agi now buffett skills, you must already have Claude Code installed on your machine. This is a developer-centric tool, so you will be working in the terminal. Follow these steps to get it running:

1. Clone the Repo: Run git clone https://github.com/agi-now/buffett-skills to get the files locally.
2. Create the Skills Folder: Inside your active project directory, create a hidden folder path: .claude/skills/buffett/.
3. Copy the Files: Move the contents of the skills/buffett folder from the cloned repo into your project's new .claude/skills/buffett/ directory.
4. Verify Installation: Start Claude Code by typing claude in your terminal. You don't need to "activate" anything manually; the system auto-discovers the SKILL.md file.
5. Trigger the Skill: Ask a question like "Evaluate the moat of [Company Name]" or "Analyze this shareholder letter." The AI will automatically pull in the Buffett reference files.

A common mistake for beginners is placing the files in the wrong directory. Claude Code only looks for skills relative to your current working directory's root. If you are in /projects/investing, the skill must be at /projects/investing/.claude/skills/buffett/SKILL.md.

Pricing Breakdown

The buffett skills Repository agi now buffett skills itself is free and open-source. There is no subscription fee to the developers of the repository. However, running this tool is significantly more expensive than standard Claude usage due to the sheer volume of data being processed.

• Repository Cost: $0 (Open Source / MIT License).
• API Token Usage: High. Because the tool reads 8+ reference files and performs multiple "thought" cycles, a single analysis can consume 30,000 to 50,000 tokens. In 2026 pricing, this could cost anywhere from $0.15 to $0.50 per deep analysis.
• Claude Code Access: Requires an active Anthropic API key or a Claude Pro subscription with credits.

For most retail investors, the cost is negligible compared to the value of the insights. For institutional users running hundreds of reports, you will need to monitor your Anthropic API dashboard closely. There are no "tiers" here; you pay for what you use in compute.

Strengths vs. Limitations

The buffett skills Repository agi now buffett skills is a specialized instrument that prioritizes analytical depth over speed. Its success depends entirely on your comfort with a command-line environment and your willingness to pay for high-token-count reasoning cycles.

Competitive Analysis

In the 2026 landscape, financial AI has bifurcated into "black-box" predictive bots and "white-box" analytical frameworks. This repository sits firmly in the latter camp, offering transparency that most commercial SaaS platforms hide behind a subscription wall.

Verdict vs. Competitors: Pick buffett skills Repository agi now buffett skills if you are a developer or a DIY investor who wants to see the "math" behind the opinion. Pick FinChat.io if you need a pretty UI and real-time price alerts. Pick Bloomberg GPT only if you are at a hedge fund with a massive budget and need proprietary terminal data integration.

FAQ

Does this tool execute trades automatically? No, it is a research and decision-support framework designed for analysis, not automated trade execution.

Do I need to be a programmer to use it? You need basic terminal proficiency to clone the repository and navigate directories, but no actual coding is required.

Does it require a monthly subscription? The repository is free and open-source, though you must pay for the Anthropic API tokens consumed during use.

Verdict with Rating

Rating: 4.4/5 Stars

The buffett skills Repository agi now buffett skills is the most rigorous value investing tool currently available for the Claude Code ecosystem. It successfully moves AI analysis beyond "summary" and into "evaluation" by forcing the model to adhere to strict margin-of-safety principles. It is an essential tool for serious value investors and analysts who already live in their terminal. However, casual investors may find the CLI setup and the $0.30-per-run token cost a barrier to entry. If you want a "buy/sell" button, look elsewhere; if you want a digital Charlie Munger to stress-test your thesis, this is it.