AI SKILLS

AI Skills & Reusable Workflows

I treat a useful AI Skill as a reusable method—not a one-off clever prompt.

My AI Skills package a repeatable way of thinking into a workflow with a defined problem, input expectations, sequence, outputs and boundaries. The objective is consistency: help the model preserve the important reasoning structure instead of improvising a different process every time.

Published September 9, 2026Shuo (Danny) Zhang

CORE QUESTIONS

The questions that matter

How is an AI Skill different from a prompt?

A prompt is an instruction for one interaction. A Skill is intended to be reused: it defines the problem, sequence, evidence handling, output structure, decision points and limits so the method survives beyond one carefully worded message.

Why do boundaries matter?

Without boundaries, a reusable workflow can become confidently inconsistent. Good Skill design says what the method is for, what it should not infer, what evidence is missing, and when the user needs another expert, source or decision process.

How do I design a Skill?

I start with the recurring human method, not the model. Then I decompose the method into stages, define what information each stage needs, specify outputs and checks, and test where the workflow loses context, compresses too early or jumps from evidence to conclusion.

What makes a Skill worth reusing?

The Skill should make the process more repeatable, transparent or efficient than starting from a blank chat. If the only advantage is a longer prompt, it probably has not yet become a reusable method.

Can an AI Skill replace expert judgment?

Usually no. A Skill can structure work, surface issues, preserve context and improve consistency, but the appropriate human decision-maker still owns consequential judgments and domain-specific conclusions.

ORIGINAL OPERATING MODEL

Method → stages → outputs → boundaries → reuse

The design principle behind my Skills is to preserve the reasoning process that matters. Project Rescue preserves diagnosis before action. No-Loss Deep Reading preserves structure and evidence before compression. From Panic to Plan preserves the distinction between facts, uncertainty and controllable next actions.

Start with the human method

Identify what a capable person actually does before writing the prompt.

Protect the information that gets lost

Find where context, evidence, uncertainty or decision logic tends to disappear.

Define stage outputs

Make each stage produce something inspectable instead of relying on invisible reasoning.

State the boundaries

Say what the Skill is not designed to decide or infer.

Package for reuse

Make the method easy to apply again without rebuilding the thinking process each time.

COLLECTION

Current public Skill families

These are the public methods currently connected to this site. Marketplace availability can change; the pages here document the method and boundaries.

Project Recovery

BOUNDARIES

A Skill is useful when the method is visible

The workflow should expose missing evidence rather than silently fill gaps.
A reusable method should preserve important context before it compresses or summarizes.
The output should make the next decision easier, not merely sound polished.
The Skill should have an explicit scope and a clear point where another source, professional or human decision is required.

Where these Skills connect

The topic hubs below show the larger systems behind the individual Skills.