# Where to Find the Key Skills Files in the ai-berkshire Repository > Locate essential key skills files within the xbtlin/ai-berkshire repository. Discover how these Markdown files power Claude and Codex command-line tools. - Repository: [Xbt Lin/ai-berkshire](https://github.com/xbtlin/ai-berkshire) - Tags: how-to-guide - Published: 2026-07-04 --- **The core skill definitions for the ai-berkshire project live as Markdown files in the top-level `skills/` directory, serving as the single source of truth for Claude Code and Codex-compatible command-line tools.** If you are searching for where the key skills files are located in the xbtlin/ai-berkshire repository, you will find them organized under the `skills/` folder at the repository root. Each Markdown file in this directory represents a distinct research workflow or automated task—such as investment research, industry analysis, or portfolio review—that powers the AI-driven research toolkit. These source files are parsed by helper scripts to generate the command-line interface and chatbot-compatible artifacts. ## Locating the Core Skills Directory The primary storage for all skill definitions is the `./skills/` path at the repository root. According to the xbtlin/ai-berkshire source code, this directory contains human-readable Markdown files that define automated research workflows. Key skill files you will find in this directory include: - **[`skills/investment-research.md`](https://github.com/xbtlin/ai-berkshire/blob/main/skills/investment-research.md)** – Company-level research workflows - **[`skills/industry-research.md`](https://github.com/xbtlin/ai-berkshire/blob/main/skills/industry-research.md)** – Sector-wide analysis frameworks - **[`skills/portfolio-review.md`](https://github.com/xbtlin/ai-berkshire/blob/main/skills/portfolio-review.md)** – Portfolio health check generation - **[`skills/earnings-review.md`](https://github.com/xbtlin/ai-berkshire/blob/main/skills/earnings-review.md)** – Structured earnings-call recaps - **[`skills/management-deep-dive.md`](https://github.com/xbtlin/ai-berkshire/blob/main/skills/management-deep-dive.md)** – Executive biography and governance analysis - **[`skills/financial-data.md`](https://github.com/xbtlin/ai-berkshire/blob/main/skills/financial-data.md)** – Financial statement fetching and normalization - **[`skills/industry-funnel.md`](https://github.com/xbtlin/ai-berkshire/blob/main/skills/industry-funnel.md)** – Industry opportunity filtering pipelines - **[`skills/quality-screen.md`](https://github.com/xbtlin/ai-berkshire/blob/main/skills/quality-screen.md)** – Quantitative screening criteria - **[`skills/investment-team.md`](https://github.com/xbtlin/ai-berkshire/blob/main/skills/investment-team.md)** – Analyst team coordination templates ## Understanding the Skill File Structure Each skill file follows a standardized Markdown format that the synchronization scripts can parse. The files contain structured sections defining the workflow's purpose, required inputs, and execution steps. A typical skill file contains sections like: ```markdown # Investment Research ## Description Gather recent filings, analyst estimates, and news for a given ticker. ## Parameters - `ticker` (string) – Stock ticker symbol ## Steps 1. Pull latest 10-K and 10-Q from SEC EDGAR. 2. Fetch consensus estimates from Bloomberg/Refinitiv. 3. Summarize key insights. ``` ## Syncing Skills with Codex Artifacts The Markdown files in `skills/` serve as the **single source of truth**, but the tooling requires generated artifacts for Claude Code and Codex compatibility. When you edit a skill file, you must regenerate these artifacts using the provided helper scripts. Run the sync script to parse the Markdown headings and create the command definitions: ```bash # Regenerate Codex-compatible artifacts from source skills python3 scripts/sync-codex-skills.py ``` Optionally, you can also run `python3 scripts/sync-codex-prompts.py` to update prompt definitions. These scripts convert the human-readable Markdown into structured files such as [`codex-skills/investment-research/SKILL.md`](https://github.com/xbtlin/ai-berkshire/blob/main/codex-skills/investment-research/SKILL.md), which expose the functionality to the command-line interface. ## Invoking Skills from the Command Line Once synchronized, skills become available as subcommands through the `ai-berkshire` CLI. You invoke a skill using its filename (minus the extension) and pass any required parameters defined in the source Markdown. Example execution: ```bash # Ensure generated Codex artifacts are up-to-date python3 scripts/sync-codex-skills.py # Run the investment-research skill (Claude-compatible) ai-berkshire investment-research --ticker AAPL ``` ## Summary - The **key skills files** in the xbtlin/ai-berkshire repository are located in the top-level `skills/` directory as Markdown files. - Each file defines a research workflow with structured sections for **Description**, **Parameters**, and **Steps**. - These source files are the **single source of truth** for the project's AI tooling and CLI commands. - After editing any skill, run `python3 scripts/sync-codex-skills.py` to regenerate the Codex-compatible artifacts. - Synchronized skills are accessible via the `ai-berkshire` CLI using the convention: `ai-berkshire `. ## Frequently Asked Questions ### What is the exact path to the skills directory in the ai-berkshire repository? The skills directory is located at the repository root under `./skills/`. This path contains all Markdown files that define the AI research workflows, such as [`investment-research.md`](https://github.com/xbtlin/ai-berkshire/blob/main/investment-research.md) and [`portfolio-review.md`](https://github.com/xbtlin/ai-berkshire/blob/main/portfolio-review.md), which serve as the source definitions for the entire toolchain. ### How do I update a skill file after editing it? After modifying any file in the `skills/` directory, you must run `python3 scripts/sync-codex-skills.py` to regenerate the Codex-compatible artifacts. The sync script parses the Markdown structure and updates the generated files in `codex-skills/` accordingly, ensuring the CLI tools reflect your changes. ### Are the skill files human-readable or machine-only formats? The skill files are **human-readable Markdown documents**. They are designed to be edited directly while containing structured sections that the sync scripts can parse to generate machine-readable command definitions for the CLI tools and chatbot interfaces. ### What is the relationship between the Markdown files and the CLI commands? The Markdown files in `skills/` serve as the **source definitions**. When you run the sync script, it converts these files into command specifications that the `ai-berkshire` CLI can execute. The CLI command names correspond to the Markdown filenames without the `.md` extension.