This repository contains local Agent Skills.

Skills are self-contained folders with instructions, scripts, and supporting assets that help agents execute recurring tasks consistently.

Each skill lives under a category directory such as skills/meta/ or skills/reporting/ and is typically centered around a SKILL.md file.

This layout is repository-local. For installed Codex skills, prefer ~/.codex/skills/.

• skill-creator: Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Codex's capabilities with specialized knowledge, workflows, or tool integrations.
• skill-installer: Install Agent Skills into $CODEX_HOME/skills from a curated list or a GitHub repo path. Use when a user asks to list installable skills, install a curated skill, or install a skill from another repo, including a private repo.

• codex-setup-report: Gather a Codex usage-consumption report from the local machine and draft a GitHub issue comment using documented current defaults plus local config checks. Use when the user reports Codex usage limits draining too fast, wants to fill a setup template for openai/codex issues, or needs a reproducible workflow for collecting Codex setup details and comparing them against the documented defaults.
• skill-review-and-auto-refactor: Audit and improve an existing Codex or OpenAI skill bundle or SKILL.md for clearer routing, stricter workflow instructions, and safer repeatable execution. Use only for existing skills, not for generic code review or net-new skill creation.

• report-generator: Generate or update a structured report from source material, then publish HTML/PDF deliverables. Use when the user asks to review a report file, analyze source inputs, write a structured report, or regenerate a styled PDF report with Pandoc and WeasyPrint.

Point Codex at this repository or copy individual skill folders into your local ~/.codex/skills/ directory, depending on how you manage skill distribution.

These skills rely on mise to provision the command-line tools they use. The repository includes a root mise.toml so the required toolchain can be installed consistently. Source code for mise is available at https://github.com/jdx/mise.

See the license information provided in each skill directory when applicable.