Awesome Skills

中文说明

My personal collection of agent skills for coding, planning, knowledge capture, slides, multi-agent work, and day-to-day setup.

I keep this collection small enough to browse, but broad enough to cover the workflows I actually reach for: making better plans, writing safer code, building richer HTML slides, saving reusable knowledge, coordinating complex agent work, and keeping my local setup repeatable.

Install the collection

npx skills add patrick-fu/awesome-skills -g

Update later:

npx skills update -g

The full collection is the easiest starting point. If one of the larger skills below is what you came for, its standalone page gives a more focused walkthrough and a one-skill install command.

Featured skills

🎞️ frontend-harness-slides

Build lively HTML slide decks with a frontend harness: stable scenes, repeatable screenshots, meaningful tests, PDF export, and online delivery.

Use it when a deck needs to look good, feel alive, and stay hard to break after many rounds of edits.

🖥️ Live demo: try the dynamic Workbench Demo. The visual styles are useful, but the bigger win is harness-backed iteration instead of fragile single-file HTML.

Standalone page: patrick-fu/frontend-harness-slides

🧭 parallel-goal-workflows

Coordinate complex work without dumping every subtask into the main conversation. It helps an agent split broad goals, run focused helpers when that is useful, review the results, and return a clean final report.

Use it for audits, research, repair loops, or any task where independent checks matter.

Standalone page: patrick-fu/parallel-goal-workflows

📚 llm-wiki-capture

Turn links, decisions, setup notes, and useful agent sessions into a Git-backed knowledge base with enough evidence to be trusted later.

Use it when something from a session should still be findable next month.

Standalone page: patrick-fu/llm-wiki-capture

Install one skill

npx skills add patrick-fu/frontend-harness-slides -g
npx skills add patrick-fu/parallel-goal-workflows -g
npx skills add patrick-fu/llm-wiki-capture -g

If you install the full collection and later install a standalone skill in the same global scope, the standalone version replaces that one local skill. It does not create a second copy.

What's inside

Slides and visual delivery

• frontend-harness-slides: HTML slide decks with style alignment, interaction, motion, screenshot checks, PDF export, and online delivery.

Knowledge and memory

• llm-wiki-capture: Save reusable source notes and session lessons into a Git-backed wiki or knowledge base.
• x-twitter-reader: Read X/Twitter posts, threads, Articles, metadata, links, and media before summarizing or archiving them.

Agent orchestration

• parallel-goal-workflows: Split complex work into owned goals, focused helper work, review, repair, and final reporting.
• claude-code-coding-agent: Use Claude Code CLI as an explicitly selected external coding agent.
• codex-coding-agent: Use local Codex CLI from another host agent or automation harness.
• cursor-coding-agent: Use Cursor CLI when Cursor is the chosen external executor.

Thinking and planning

• brainstorm: Explore a rough idea before turning it into a plan.
• explore-and-plan: Turn a mostly chosen direction into executable steps and acceptance checks.
• faster-learning-coach: Convert explanations into short learning loops with practice and review.

Engineering workflow

• write-unit-test: Write or improve unit tests around behavior, regressions, and domain rules.
• log-driven-debugging: Add targeted logs, rerun the real scenario, and trace the first useful divergence.
• generate-commit-message: Draft a concise English commit message with simple formatting constraints.

Personal setup

• home-config-sync: Manage a personal ~/.dotfiles bare-repo workflow, including first setup, new-machine deploys, pull safety, and pushes.