agentleFS
Sign inStart free
CipherNovaaCipherNovaa/agent-skills/skills/tools/learn-accelerator/SKILL.md

learn-accelerator

Skill0 starsChanged 12 days agoMITView on GitHub

Explain a math problem, word problem, or programming concept in the user's own study style (answer-first, step-by-step, a memory trick, a worked example or gotcha). Use when the user asks to solve, explain, understand or learn a math or programming topic.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    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:

    Call the native learn tool (provided by the learn-accelerator plugin), passing the question verbatim as topic:

    line 35
  2. Distinctive:

    Reply with the tool output verbatim: Do not summarize it, comment on it, or add your own explanation.

    line 46
  3. Distinctive:

    If the tool returns an error (e.g.

    line 50
  4. Distinctive:

    When the user states a preference ("more worked examples", "skip the code", "shorter"), append one short bullet to the learning-style …

    line 53

≈ “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

learn-accelerator

A personal "learning accelerator". It explains math problems, word problems and programming concepts in the user's study style, and learns that style over time. The same engine can sit behind a handwriting-tablet //learn command, so desktop, chat and tablet give consistent answers.

When to Use

Activate when the user wants to understand or learn, not just get a one-line answer:

  • A math problem or expression to solve or evaluate ("solve 4^2 + (15/3)*2 - 7").
  • A word problem.
  • "Explain / teach me / help me understand / how does X work" for a math or programming concept ("what's a pointer in C++", "why does multiplication come before addition").
  • Anything that looks like studying, homework or concept review.

For a quick factual lookup with no teaching needed, answer directly; this skill is for explanations that should stick.

Procedure

  1. Call the native learn tool (provided by the learn-accelerator plugin), passing the question verbatim as topic:
    learn(topic="<the user's question, verbatim>")
  2. web is optional: "auto" (default) fact-checks concepts against the local SearXNG and skips bare arithmetic; pass "off" only for the fastest answer.
  3. The tool runs fully local and synchronously and returns the finished explanation. It already loads the learning-style profile (MYOS_LEARNING_STYLE, default ~/.hermes/learning_style.md) and samples two of the user's own well-structured math notes as few-shot examples.
  4. Reply with the tool output verbatim. Do not summarize it, comment on it, or add your own explanation.

Pitfalls

  • If the tool returns an error (e.g. the local model is unreachable), say so plainly. Never write an explanation yourself instead — the point is the personalised, fact-checked engine, not the chat model's guess.
  • When the user states a preference ("more worked examples", "skip the code", "shorter"), append one short bullet to the learning-style profile with a file tool. The engine reads it on every run, so the style grows with the user.

Verification

  • The reply is the tool's markdown, unchanged: answer first, a numbered step-by-step, a "🧠 Memory trick" line, and a worked example or gotcha.

Cadence

On demand.

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.

Reviews can't be read right now.

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.

Check your setup, free