Hung-Yi Lee Teaching Style

What it does

When your agent uses it

• Learning how transformers or attention work, with intuition before math
• Getting a roadmap-first overview of LLMs or AI agents
• Preparing a lecture on a machine learning topic with a story-like flow
• Explaining speech modeling concepts with everyday analogies

Example prompts

• “Explain self-attention the way Hung-Yi Lee would, starting from the problem it solves.”
• “Give me a roadmap-first overview of how LLM agents work, with an everyday analogy.”
• “Walk me through why deep learning needs more than linear regression, tracing how the idea evolved.”
• “Help me work out why my training loss does not drop, explained in a teaching style.”

Workflow steps

12 steps, taken from the step headings in SKILL.md.

1. 1Start With A One-Sentence Punch「一言以蔽之」
2. 2Black Box Before Internals
3. 3Anticipate Confusion And Surface It「你可能會想說…」
4. 4Problem Before Method「怎麼辦呢？」
5. 5Concrete Example Immediately「比如說…」
6. 6Restate The Same Idea From Multiple Angles
7. 7Scale And Surprise「你知道嗎…」
8. 8Honest Scope Markers「先抓核心」
9. 9Vivid Analogy — Concrete, Apt, Often Anime
10. 10Guided Discovery「你會發現…」
11. 11Transition-Rich Flow
12. 12Warm Ending With Recap

What it can do on your machine

Read from SKILL.md and the folder at commit df5665d. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Ships 1 file in scripts/, which the agent can run.

  Shell commands in SKILL.md call:

  • python3

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Hung-Yi Lee Teaching Style loads about 13k tokens when it runs, and up to ~33k if it reads all its reference files. Until then it costs about 95 tokens; SKILL.md has 6,923 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~95

When it runs · the whole SKILL.md, loaded when a task matches

~13k

With references · SKILL.md plus every file in references/, read only if the agent opens them

~33k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

Without a licence we can't republish the file, so here is its outline and opening line. It has 6,923 words (~12,937 tokens).

name

hung-yi-lee

Read the full SKILL.md on GitHub

Files

SKILL.md and 279 other files (scripts, references, assets) in the repository root of voidful/hung-yi-lee-skill.

• SKILL.md
• .gitignore
• AGENTS.md
• README.en.md
• README.md
• agents/openai.yaml
• assets/hung-yi-lee.png
• outputs/query-briefs/什麼是-context-engineering.md
• outputs/query-briefs/什麼是-transformer-attention.md
• raw/youtube/channel_videos.json
• raw/youtube/transcript_index.json
• raw/youtube/transcripts/2rcJdFuNbZQ.md
• raw/youtube/transcripts/8iFvM7WUUs8.md
• raw/youtube/transcripts/9HPsz7F0mJg.md
• … and 266 more

Open the folder on GitHubat commit df5665d