Skip to content
Agentmart
Submit a tool

explain

In an lstack learning vault (a folder containing lstack.yaml), explain one node from its sources in plain, intuition, or formal mode, pausing for recall…

Skill by sscw32

IndexedMIT
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/sscw32-explain

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

In an lstack learning vault (a folder containing lstack.yaml), explain one node from its sources in plain, intuition, or formal mode, pausing for recall questions and ending with one retrieval question. Use for "/explain", "explain kernel", "teach me X", "walk me through this node", "I don't get Y". Skip when the folder has no lstack.yaml, when the user wants to be tested (/quiz, /drill), or when they ask to explain code or an error message.

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/explain

What's inside

Skill name
explain
Description
In an lstack learning vault (a folder containing lstack.yaml), explain one node from its sources in plain, intuition, or formal mode, pausing for recall questions and ending with one retrieval question. Use for "/explain", "explain kernel", "teach me X", "walk me through this node", "I don't get Y". Skip when the folder has no lstack.yaml, when the user wants to be tested (/quiz, /drill), or when they ask to explain code or an error message.
License
MIT
Instructions
51 lines in SKILL.md · view on GitHub

Files (2)

  • SKILL.md
  • openai.yaml · agents/

SKILL.md (first lines)

# Explain

Teach one node at a time, from its sources, with retrieval built in. Mark the node as studied only after a yes.

## Preconditions

Confirm `lstack.yaml` exists in the working directory or a parent. If not, say this isn't a vault and suggest `/setup-vault`. Read `MISSION.md`, `kb/today.md`, and the newest file in `log/`. Open with one line on where the user is.

## Argument

The text after the skill name in the user's message is a node id or a phrase, plus an optional mode `plain`, `intuition`, or `formal`. Resolve a phrase to a node by title or `## Terms`. If nothing matches, say "no node yet" and offer `/add-source`. If there is no argument, ask. Default mode: `plain`, or `formal` when `lstack.yaml` voice has `verbosity: thorough` and `register: formal`.

## Steps

1. Read the node file, its `sources` entries in `sources/`, and its prereqs' Summaries. Do not explain from memory alone when a source exists; cite the source section you used.
2. Explain in at most about five paragraphs.
   - `plain`: the idea in everyday words, one worked example, one common confusion.
   - `intuition`: lead with the node's `## Analogies` entry (or propose one), then "when would you reach for this", then the idea.
   - `formal`: the source's definitions and notation, one theorem or property with its conditions stated, one example.
3. Every three to five paragraphs stop and ask one recall or self-explanation question ("say back the definition", "why did that step need the prereq?"). Wait for the answer. Correct it if wrong; never confirm a wrong answer.
4. End with exactly one retrieval question about the node. Wait for the answer and judge it. This is not a test for status; say so if asked.
5. Propose the node edit, preview, yes, write:
   - `status: new → learning` if the node was `new`.
   - `last_studied: <old> → <today>`.
   - `sessions`: append today if not present.
   - Any analogy, term, or heuristic that came up and is worth keeping, as `+` body lines under the right section.
6. Run `node .lstack/lstack.mjs build`, then `node .lstack/lstack.mjs lint`. Report one line. If `lstack.yaml` says `node_available: false`, rebuild `kb/index.md` and `kb/today.md` by hand and say so.

## Preview format

```
### Proposed write 1 of 1: kb/linear-maps/kernel.md
**Frontmatter**
- status: new → learning
- last_studied: null → 2026-09-15
- sessions: [] → [2026-09-15]
**Body**
+ ## Analogies
+ - The kernel is what the map flattens to a point.
```

Reply **yes** to write, **no** to skip, or name what to change.

## Rules

- Never change `status` beyond `new → learning`, and never touch `confidence`, `last_tested`, `last_result`, `next_review`, or `passes_in_a_row`.
- Praise the work, not the person.
- Never write into `sources/`, `notes/`, or `problems/attempts/`. Never edit generated files.
- No hooks, no sub-agents, no argument substitution.

From sscw32/lstack, MIT.

More from sscw32/lstack

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
Sep 15, 2026 (1 month ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from sscw32

Cite this page

explain skill. Agent Mart. https://agentmart.dev/skills/sscw32-explain (last updated 2026-10-24)