agentleFS
masbrokemanaajamasbrokemanaaja/reethink/skills/plain-technical/SKILL.md

plain-technical

Skill3 starsChanged 55 days agoMITView on GitHub

Makes technical explanations readable by a non-specialist without making them less technical. Says what a thing does with a subject that acts, explains a term in the same breath it first appears, and grounds every abstract claim in a concrete value, while keeping the real names, numbers and identifiers that let the reader search and verify. Use whenever explaining what code does, why a bug happened, what a change affects, what an error means, or how a system works, and whenever the audience may include someone who does not work in that stack. Also use when the user says "explain simply", "in plain English", "I do not follow", "jelasin yang gampang", "bahasanya kaku", "biar awam paham", or asks what something means.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 6 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Write so a smart person outside this stack can follow it, and a specialist inside it still trusts it.

    line 24
  2. Distinctive:

    If a sentence's main noun ends in -tion, -ment or -ance, look for the verb hiding inside it.

    line 42
  3. Distinctive:

    Abstract claims with nothing concrete attached: Every claim about behaviour gets a value, a path, a duration or an example input.

    line 49
  4. Distinctive:

    Short verbs over long ones: Use, not utilize. Start, not initiate.

    line 64
  5. Distinctive:

    When you cannot tell, explain the term once in brackets.

    line 92
  6. Distinctive:

    Keep technical terms in the form practitioners actually use in that language, which is usually the English term.

    line 97

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Plain Technical

Write so a smart person outside this stack can follow it, and a specialist inside it still trusts it. Those are the same piece of writing, not two.

Plain does not mean simplified. Every real name, version, number, file path and error string stays. The reader needs them to search for the thing, to check you, and to talk to someone else about it. Take those away and the text is easier to read and useless.

This is about how it reads, not how long it is. A clear explanation is sometimes longer than the stiff one, because a term got defined or an example got added. That extra sentence carries information, so it is not padding.

What makes writing stiff

Five patterns account for most of it. Each has a mechanical fix.

  1. Nouns doing the work of verbs. "Implementation of origin validation has been applied" becomes "the check rejects requests from other sites". If a sentence's main noun ends in -tion, -ment or -ance, look for the verb hiding inside it.
  2. Sentences with no actor. "An error was encountered" becomes "the build failed because X". Name who or what did it.
  3. A term used but never explained. The first time a term appears, give its meaning in the same sentence, usually in a clause or in brackets. Once is enough; after that, use it freely.
  4. Abstract claims with nothing concrete attached. Every claim about behaviour gets a value, a path, a duration or an example input.
  5. Corporate register. Utilize, leverage, facilitate, robust, seamless, in order to, it is important to note that. Use, use, help, and delete the rest.

How to say it instead

  • Lead with what it does, in the order it happens. A request arrives, the check reads where it came from, it does not match, it is refused.
  • Define on first use, then move on. "It stops at the coast because land cells have zero speed (the wave cannot travel through them)."
  • One analogy, and only for a mechanism the reader has no model for. The analogy is a handle, not evidence. Never make a claim inside the analogy: the claim comes from the measurement, the analogy only helps the reader hold it.
  • Short verbs over long ones. Use, not utilize. Start, not initiate. Show, not surface.
  • Name the consequence the reader cares about. Not "the cache TTL is 300 s", but "a change takes up to 5 minutes to show up for users, because the cache holds it for 300 s".

What must stay

Removing these makes the text simpler and worse:

  • The real term. Say it plainly and name it: "it spreads outward one cell at a time (this is the eikonal solver)". The non-specialist now has a word to search for, and the specialist knows exactly what you built.
  • Numbers, never adjectives in their place. "22 km bands in the wave" rather than "some visual artefacts".
  • Identifiers: file and line, version, flag, environment variable, exact error text. These are what the reader acts on.
  • The honesty markers: measured, assumed, untested, and what you did not check. Making prose friendlier must never make a claim stronger than the evidence behind it.

Read the audience from their own words

The person's vocabulary tells you the level. Someone who writes "the pod keeps OOMKilling" does not need containers explained. Someone who writes "the app crashes on the server" does. Explaining what the reader already knows is its own kind of noise, and it reads as condescension.

When you cannot tell, explain the term once in brackets. It costs the specialist a second and saves everyone else the question.

Replying in a language other than English

Keep technical terms in the form practitioners actually use in that language, which is usually the English term. Forcing a translation nobody says makes the text harder to read, not easier, and it breaks search. Explain the concept in the local language; keep the term as it is spoken.

Self-check before sending

  • Could someone outside this stack say back what changed and why it matters?
  • Is every term either explained once or genuinely common ground?
  • Does each claim about behaviour have a number, a path or an example on it?
  • Did making it friendlier drop a caveat, a measurement, or a real name? Put it back.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.