agentleFS
agentcrew-academyagentcrew-academy/harness-starter-kit/skills/explain/SKILL.md

explain

Skill41 starsChanged 18 days agoMITView on GitHub

Re-explain the previous reply or a specified topic in plain language a high-schooler could follow, no jargon. Triggers: /explain, plain English, I don't get it, ELI5, too much jargon.

122 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use debug-fix, root-cause and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • zebbernSkill

    debug-fix

    Find and fix a bug or issue — from any source (GitHub issue, error message, user report, or observed behavior)

    Also in 3 projects that use this

  • Ingramradical235Skill

    root-cause

    Structured root cause analysis using WHERE (MECE decomposition) then WHY (causal structure diagrams) pipeline. Includes coin-flip detection, logical-leap detection, and vicious cycle intervention. Use when buffer enters YELLOW/RED, when metrics stall, or when Actions execute but don't produce results.

    Also in 5 projects that use this

  • ralphkraussSkill

    resolve-conflicts

    Resolve git merge conflicts or re-apply changes from a stale branch onto current main. Use when user says "resolve conflicts", "merge conflicts", "rebase branch", "branch is behind", "re-apply changes", "port branch", or a branch is too far behind main to merge cleanly.

    Also in 4 projects that use this

  • neurofooSkill

    socratic

    Socratic questioning to examine beliefs, uncover assumptions, and develop deeper understanding. Use to challenge thinking, evaluate proposals, or teach without lecturing.

    Also in 4 projects that use this

  • anombyte93Skill

    discover

    Phase 1 of the prd-taskmaster pipeline: brainstorm-driven discovery. Delegates to superpowers:brainstorming in Interactive Mode (one adaptive question at a time), or self-brainstorms in Autonomous Mode when no user is present. Intercepts before the brainstorming chain hands off to writing-plans — this skill owns the exit. Extracts constraints, calibrates scale (Solo / Team / Enterprise), and advances the pipeline to GENERATE.

    Also in 6 projects that use this

  • mitsuhikoSkill

    summarize

    Fetch a URL or convert a local file (PDF/DOCX/HTML/etc.) into Markdown using `uvx markitdown`, optionally it can summarize

    Also in 8 projects that use this

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 5 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 109 edited versions we read keep these.

  • 5

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    When someone types /explain (with or without a topic), re-explain the subject to a standard of "a high-schooler with no background in this …

    line 8
  2. Distinctive:

    If it's unclear which part is meant, pick the part the user is most likely stuck on, and open with "I'm guessing you mean X — let me know …

    line 15
  3. Distinctive:

    If you do, call it with a prompt that pastes in the full "Hard rules" section below verbatim (the other process can't see this file, so it …

    line 24
  4. Distinctive:

    Only explain it the first time it shows up.

    line 51
  5. Distinctive:

    Don't drop jargon like "SOW," "addendum," "roster," "pipeline," or "enterprise" without explaining it — either use an everyday word …

    line 52
  6. Distinctive:

    Going deeper: if there's a more detailed layer, cover it briefly in a last short section so the reader can choose to read on.

    line 66

“Common” is measured: kept by at least 6 in 10 of the 109 edited versions of this file we could read. “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

Explain — Say It So a Layperson Gets It

When someone types /explain (with or without a topic), re-explain the subject to a standard of "a high-schooler with no background in this could follow it right away."

What to explain

  • No topic given (just /explain): explain my own previous reply — whatever the user didn't follow is that.
  • Topic given (/explain how the quote was calculated): explain that specific thing.
  • File path given (/explain <file path>): read that file first, then re-explain it following the rules below.
  • If it's unclear which part is meant, pick the part the user is most likely stuck on, and open with "I'm guessing you mean X — let me know if that's not it."

Document mode: if you have a way to delegate to another model, prefer that when the target is a file path

Skip this section when the target is the previous reply or an abstract topic with no matching file — a separate process can't see this conversation.

It only applies when the target is a specific file path:

  1. This assumes you have some CLI tool that can run a headless prompt against another model and read a local file by absolute path (this section was originally written around agy, a wrapper for Google's Antigravity CLI — swap in whatever equivalent tool you actually have, or skip this entirely if you don't have one).
  2. If you do, call it with a prompt that pastes in the full "Hard rules" section below verbatim (the other process can't see this file, so it needs the whole ruleset spelled out), plus the absolute path to read. For example:
    Re-explain this document in plain [language] a high-schooler could follow. Rules:
    1. No mixing languages; a genuinely untranslatable term can stay in the
       original, with a parenthetical plain-language explanation the first
       time it appears.
    2. No jargon, no abbreviations.
    3. Short sentences, one idea each.
    4. Lean on comparisons and everyday examples.
    5. Lead with the conclusion / what it means for the reader, details after.
    6. Say less and say it clearly -- don't cram in every detail.
    Structure: one-sentence summary -> plain-language breakdown (2-4 points) ->
    what it means for you -> a closing offer to go deeper if they want it.
    Read this file: <absolute path>
  3. Fall back to writing it yourself (never hand over a rule-violating output as-is) whenever any of these hold:
    • the tool isn't installed / command not found
    • non-zero exit code, timeout, or empty output
    • the output still breaks a "Hard rules" item below (mixed languages without a parenthetical, untranslated jargon, and so on)
  4. Whether the draft came from the tool or you wrote it yourself, always run the "Self-check" below before sending it.

This is optional acceleration, not a requirement — everything else in this file works the same with or without a second CLI in the loop.

Hard rules (breaking these means the explanation failed)

  1. Reply in the language the user is using, and re-explain it in plain language a high-schooler could follow — no jargon.
    • If a technical term genuinely has no good equivalent, or the original term is actually clearer than any translation (e.g., API, SaaS, PDF), you can keep the original term — but immediately follow it with a plain-language explanation in parentheses the first time it appears. Example: "API (a way for two pieces of software to talk to each other)."
    • Only explain it the first time it shows up; after that you can use it directly.
    • Don't drop jargon like "SOW," "addendum," "roster," "pipeline," or "enterprise" without explaining it — either use an everyday word instead, or add a parenthetical explanation.
    • Names of people and companies are the exception: keep them exactly as written. Don't translate, transliterate, or add titles ("John" stays "John", not "Manager John"); a changed name no longer matches the person the reader knows.
  2. No jargon, no abbreviations. Say "spreadsheet" instead of "workbook file," say "pay part of it upfront" instead of "remit an initial installment."
  3. Short sentences. One idea per sentence — don't cram three concepts into one sentence.
  4. Use comparisons and everyday examples. Give an everyday analogy before anything abstract ("it's like...").
  5. Lead with the conclusion / what it means for the reader, then the details. What the user wants to know first is "what does this mean for me — good news or bad."
  6. Don't pile on information. Say less and say it clearly, rather than cramming in every detail. When something doesn't land, the usual cause is "too much," not "too little."
  7. No emoji by default (unless the existing conversation is already using them).

Suggested structure (adapt freely — not every explanation needs all four)

  1. One-sentence summary: what this actually is / what happened.
  2. Plain-language breakdown: 2–4 points, each with a comparison or example.
  3. What it means for you: good news or bad, whether you need to do anything, whether it costs money, any risk.
  4. Going deeper: if there's a more detailed layer, cover it briefly in a last short section so the reader can choose to read on. Don't end by asking "want me to go on?" — finish the explanation in one reply.

Self-check (run through before sending)

  • Any jargon or foreign-language term left unexplained? → Fix it. (Names of people and companies don't count; changing a name is the error.)
  • Any sentence that requires understanding another term first? → Break it apart and explain that first.
  • Could a high-schooler read this and repeat it back in their own words? → If not, simplify further.

Before / after example

Before (fails the bar):

I converted the SOW Addendum to a PDF and sent it to Sho — it's the quotation for the recorded course, with the fee folded into the Stage 2 balance.

After (passes the bar):

I put together a supplementary contract for you, saved it as a file that's ready to print and sign, and sent it to your client Sho. It's about a pre-recorded online course you're offering him as an add-on — the document just spells out how much that course costs. He won't be billed separately for it; it gets added to the balance he still owes from before.

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.