An agentic skills framework & software development methodology that works.
Agent Skills pack for Claude Code, Codex, and Cursor
This repo packages five reusable Agent Skills for editing, teaching, clarifying, shortening, and visual explanations. Each skill lives in its own Markdown file and is meant to be triggered with the agent’s normal skill syntax, such as `$teach` or `/teach`.
Builders who want reusable prompt skills for Claude Code, Codex, or Cursor.
You can turn rough notes into clearer prompts, cleaner writing, lessons, shorter instructions, or visual explanations without rewriting your workflow.
What it does
Deslop skill
Removes AI writing patterns while keeping meaning and voice for replies, emails, docs, articles, lessons, scripts, and business writing.
Teach skill
Turns ideas into explanations or lessons for course material, technical articles, video scripts, workshops, newsletters, and guides.
Clarify skill
Converts a rough request into a prompt a fresh agent can execute, with scope and success criteria.
Compress skill
Shortens prompts, specs, plans, skills, and notes while preserving requirements.
Explain visually skill
Builds a responsive HTML page with writing and diagrams for architectures, flows, changes, comparisons, and state transitions.
Validation workflow
Includes automated checks and a local `make validate` command to catch frontmatter, metadata, link, and repo drift issues.
How to get it
- 1Browse the collection and choose where to install it
npx skills@latest add owainlewis/agent-skills
- 2Install one skill directly
npx skills@latest add owainlewis/agent-skills --skill teach
- 3Update installed skills with
npx skills@latest update
README
Agent Skills
Five skills for prompts, writing, and teaching.
The repository contains five skills. Each solves a different problem and works with agents that support the Agent Skills format.
Install
Browse the collection and choose where to install it:
npx skills@latest add owainlewis/agent-skills
Install one skill directly:
npx skills@latest add owainlewis/agent-skills --skill teach
Add --global to install for your user account instead of the current project. The CLI supports Codex, Claude Code, Cursor, and other clients that implement Agent Skills.
Update installed skills with:
npx skills@latest update
The skills
| Skill | Job | Use it for |
|---|---|---|
deslop | Remove AI writing patterns while preserving meaning and voice. | Replies, emails, documentation, articles, lessons, scripts, and business writing. |
teach | Explain how something works or teach the reader how to do it. | Course lessons, technical articles, video scripts, workshops, newsletters, and standalone guides. |
clarify | Turn a rough request into a prompt a fresh agent can execute. | Voice dumps, vague plans, incomplete tasks, and prompts that need scope or success criteria. |
compress | Shorten information or instructions without losing meaning or requirements. | Prompts, specifications, plans, skills, notes, and brain dumps that cost too many tokens. |
explain-visually | Build a responsive HTML page with clear writing and diagrams. | Architectures, flows, changes, comparisons, state transitions, and technical concepts. |
Invoke a skill using your agent's normal syntax, such as $teach in Codex or /teach in Claude Code.
Example prompts
Use $deslop to edit this article without losing my dry tone.
Use $teach to turn these notes into a course lesson on how tool calling works.
Use $clarify to turn this voice dump into a prompt for a coding agent.
Use $compress to cut this specification to the smallest version that preserves every requirement.
Use $explain-visually to show how this request moves through the system.
How they fit together
Use clarify to write the prompt. Use teach to create the tutorial. Use explain-visually when it needs diagrams. Use deslop to remove AI writing patterns. Use compress to shorten instructions without losing requirements.
You can use one skill alone or combine them. A typical tutorial workflow is clarify -> teach -> deslop. Add explain-visually when the tutorial needs a diagram.
Design principles
- One skill, one recurring job.
- Plain descriptions that make triggering predictable.
- Instructions that preserve facts, constraints, and user intent.
- Portable Markdown with no required service or framework.
- Examples only when they define behaviour or expose an edge case.
- Automated checks for frontmatter, metadata, links, and repository drift.
Development
Each skill lives in skills/<name>/SKILL.md and follows the Agent Skills specification. Codex UI metadata lives beside it in skills/<name>/agents/openai.yaml.
Run the local checks before committing:
make validate
See CONTRIBUTING.md for the quality bar and contribution workflow.
Software delivery skills such as spec, plan, implement, and task-to-pr live in owainlewis/blueprint.
Credits
deslop is adapted from Lauren Tan's Unslop.
teach is informed by Lauren Tan's Teach, rewritten here for tutorial, course, article, video, and workshop production.
License
Files in the repo
- .github
- scripts
- skills
- .gitignore
- AGENTS.md
- CONTRIBUTING.md
- LICENSE
- Makefile
- README.md
Discussion (0)
Ask about usage, or say what you built with itSign in to join the discussion.
No comments yet. Be the first to say what this is good for.
More skills

Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local deterministic AST parsing, every edge explained, no vector store.
Topic in, narrated explainer video out. A Claude Code / Codex skill that turns any topic into a black-canvas motion-graphics explainer video with TTS voiceover, subtitles and a chapter progress bar. Chinese or English; every frame drawn in code with Remotion.
Public repository for Agent Skills
Open-source AI job search: scan job portals, evaluate listings into a structured A-H report with a global 1-5 score, tailor your CV, track applications — runs locally in your AI coding CLI (Claude Code, Codex, OpenCode, Antigravity…)

Production-grade engineering skills for AI coding agents.