Part 2 · The Harness Around the Model

The Essence of Skill

Good loop vs bad loop upgraded; how Skill changes an Agent's execution path

THE QUESTION THIS PAGE ANSWERS

ANSWER FIRST

What is the key idea behind “The Essence of Skill”?

Good loop vs bad loop upgraded; how Skill changes an Agent's execution path

DECISION RULE

Make the claim earn its place. Use this page as a decision aid, not a definition to memorize. Connect the idea to one real task, one observable result, and one failure that would change your mind.

TRY NEXT

Write one question you could answer with evidence after trying this idea.

WATCH FOR

A conclusion that sounds complete but leaves the key assumption untested.

User Command: "Deploy a release" — same task, drastically different execution processes for the two Agents on left and right
Comparative Simulation: Click the "Simulate" button on each side
Inefficient Agent without Skill
Rounds 0
Tokens 0
Errors 0
Every step relies on general capabilities
Efficient Agent with Skill
Rounds 0
Tokens 0
Errors 0
Executes step by step per Skill manual
Skill Definition Card and Product Decisions
📋
Skill Definition Card
One SKILL.md file = one skill
Skill Name release
Triggers
deploy build release publish
Steps
1Run tests 2Build 3Update version 4Tag 5Push remote
Allowed Tools
run_command edit_file git
Storage path: ~/.the example system/skills/release/SKILL.md or {project}/.the example system/skills/
📌 Skills are encapsulated experience: Write senior engineers' workflows into a manual the Agent can understand. The Agent no longer starts from scratch every time — it stands on the shoulders of those who came before.
📌 Efficiency gap can exceed 3×: Without Skills, an Agent fumbles with general capabilities; with Skills, it executes best practices. Token consumption differs by 3.75×, errors go from 3 to 0.
📌 Questions PMs need to answer: Which high-frequency operations get pre-installed Skills? Can users create their own Skills? How to handle Skill conflicts? Who maintains and updates Skills?
Takeaway
Takeaway Skills turn "fumbling" into "execution." For the same release task: without a Skill, the Agent makes 12 trial-and-error rounds using 15K Tokens and 3 errors; with a Skill, done in 5 rounds using only 4K Tokens and zero errors. Skills are the experience-transmission mechanism for Agents.

How “Comparative Simulation: Click the "Simulate" button on each side” changes an answer

“Good loop vs bad loop upgraded;” shows that a model does not process the “word count” we see. It processes Token pieces. Tokenization affects input length, how much context fits, and how much computation a request consumes.

Length, information, and context are different

As “Good loop vs bad loop upgraded;” grows, separate three questions: how many Tokens the text becomes, which pieces can change the current decision, and whether older material has fallen outside the context window. Removing repetition is often more useful than simply making the window larger.

Keep what can change the decision

Use “Good loop vs bad loop upgraded;” as an A/B test: keep the same question while removing repeated background, compressing format, and trimming irrelevant history. Compare answer quality, latency, and Token count.

From “Comparative Simulation: Click the "Simulate" button on each side” to “Skill Definition Card and Product Decisions”

“Comparative Simulation: Click the "Simulate" button on each side” grounds the problem in “Inefficient Agent without Skill Rounds 0 Tokens 0 Errors 0 ▶ Simulate Reset Every step relies on general capabilities Efficient Agent with Skill Rounds 0 Tokens 0 Errors 0 ▶ Simulate Reset Executes step by step…”. “Skill Definition Card and Product Decisions” then moves it toward “📋 Skill Definition Card One SKILL.md file = one skill Skill Name release Triggers deploy build release publish Steps 1 Run tests → 2 Build → 3 Update version → 4 Tag → 5 Push remote Allowed Tools run_command e…”. Together, they show that the lesson is not just a conclusion to remember, but a claim with conditions.

Carry the judgment into the next situation

For long text, keep what can change the conclusion before compressing format and history. A larger context is worth its cost only when the added information is useful.

  • “Comparative Simulation: Click the "Simulate" button on each side”: Inefficient Agent without Skill Rounds 0 Tokens 0 Errors 0 ▶ Simulate Reset Every step relies on general capabilities Efficient Agent with Skill Rounds 0 Tokens 0 Errors 0 ▶ Simulate Reset Executes step by step…
  • “Skill Definition Card and Product Decisions”: 📋 Skill Definition Card One SKILL.md file = one skill Skill Name release Triggers deploy build release publish Steps 1 Run tests → 2 Build → 3 Update version → 4 Tag → 5 Push remote Allowed Tools run_command e…

The final “Finish by testing the claim” brings the discussion to “Good loop vs bad loop upgraded”. The useful thing to carry forward is knowing which judgments must be revisited when input, scale, or risk changes.

Mark as learned Your reading progress updates automatically
← PreviousNext →

Keep reading

The next useful article in the thread.

ARTICLE DISCUSSION

Leave one useful thought here.

Keep the idea that clicked, the question that stayed open, or a small note for the next learner.

Discussing The Essence of Skill The Harness Around the Model
3discussionsArticle discussion · synced with the Circle
View in the learning circle
AM
Asha MorganContent editor
INSIGHTField note

I turned one judgment from this article into a small experiment I could run today. Knowing what to observe next is more useful than simply remembering the conclusion.

ARTICLE DISCUSSION7 helpful
LH
Lin HarperIndie developer
INSIGHTInsight

After reading this, I first looked for the conditions behind the idea instead of copying the method into a project. That order made the later trade-offs much clearer.

ARTICLE DISCUSSION5 helpful
KM
Kiki MooreProduct operations
QUESTIONQuestion

When this judgment reaches real work, which constraint should be added first? I am curious which step matters most between reading and the first practical attempt.

ARTICLE DISCUSSION4 helpful