nweii/agent-stuff

save-term

Save a term, phrase, or coined concept as a Term note in the Brain vault under 02-Evergreen/Terms/.

First seen May 12, 2026

Installation

$ npx skills add nweii/agent-stuff --skill save-term

Summary

  • Save a term, phrase, or coined concept as a Term note in the Brain vault under 02-Evergreen/Terms/.
  • Two modes: capturing a nice term encountered in reading (with source quote) or recording one coined during conversation (with brief gloss).
  • Triggers: '/save-term <term>', 'save this term', 'add this to my terms'.

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from nweii/agent-stuff · top by installs.

npx skills add nweii/agent-stuff

Browse all from nweii/agent-stuff

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Repository health

Stars 9
License LICENSE
Default branch master
Status Active

Skill metadata

Parsed from SKILL.md frontmatter.

Version1.0.1
Declared agents claude-code
More metadata
author
nweii
version
1.0.1
internal
1

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 6,871 B
  • docs SUMMARY.md 329 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 9 installs

SKILL.md

Save term

Save a term or phrase as a Term note in 02-Evergreen/Terms/. Two modes:

  1. Encountered — a term lifted from a book, article, conversation, or other external source. Body anchors on a source citation + quote.
  2. Coined — a term that emerged during conversation (often a compact replacement for a wordy phrase, or a fresh concept the user wants to lock in). Body is a brief gloss with a usage example or context.

This skill is not for general note-taking. It writes one specific shape of note — short, atomic, and aligned with the existing vault Term convention. For broader capture, use save-napkin-note.

Defer to AGENTS.md / CLAUDE.md

Vault layout, metadata semantics, and tool-choice rules live in the vault's AGENTS.md and CLAUDE.md. This skill only adds the procedure for writing one note type. When unsure about a field's meaning or which tool to use for vault writes, read those first.

Arguments

  • $ARGUMENTS: the term or phrase to save. If omitted, ask. If the term contains punctuation tricky for filenames (/, :, etc.), confirm a filesystem-safe variant before writing.

Process

1. Determine mode

Ask if not obvious from context:

  • Encountered — user is referencing something external (a book, article, person they're quoting, web clipping). Need a source.
  • Coined — emerged in this conversation, no external source. Likely came out of a phrase-revision pass or a fresh framing.

2. Gather what the body needs

For encountered terms:

  • Source citation — wikilink to an existing vault note if possible ([[Source note title]]), otherwise a markdown link to the URL with the article/book title as link text. Check 04-Entities/ and 03-Records/Clippings/ for an existing source note before defaulting to a raw URL.
  • Quote — verbatim passage from the source that defines or illustrates the term. Use a blockquote (>). Don't paraphrase.
  • Optional: brief framing sentence above the quote if needed for context, though most existing Term notes go straight to the quote.

For coined terms:

  • Gloss — 1–2 sentences defining the term in plain language. The user's own wording is preferable to your phrasing — pull from the conversation if a clean definition already exists there.
  • Usage example or context — optionally, the wordy phrase it replaces (if it came from a phrase-revision pass) or a sentence showing it in use.
  • Optional source — if it was inspired by something external, link or cite it lightly.

3. Pick an icon (optional but encouraged)

Existing Term notes usually have a Lucide icon in frontmatter (e.g., icon: LiFilter, icon: LiWaves). If you can suggest a fitting one, do — otherwise leave the field blank rather than guessing badly. The suggest-lucide-icons skill is the right tool when the user wants help picking one.

4. Find related notes

Search the vault briefly for adjacent Term notes or concept notes (02-Evergreen/) that this term connects to. Two or three is plenty — these populate ## See also. Don't force connections; if nothing obvious comes up, leave the section empty or omit it.

5. Write the note

Path: 02-Evergreen/Terms/{term}.md

Filename casing: match the user's preference if they specified one; otherwise default to sentence case for multi-word terms (Bottleneck optimization.md) and lowercase for single common-noun terms (burnout.md). Existing examples are mixed — don't agonize over it.

Filename collision: before writing, check if the file exists. If it does, surface to the user — don't silently overwrite. Options: append to the existing note, pick a more specific filename, or abort.

Frontmatter shape (matching existing Term notes):

---
icon: Li{LucideIconName}     # optional
publish: false
related:                      # optional, omit if empty
  - "[[Adjacent term]]"
tags:
  - {topical tags}
  - evergreen
  - terms
created: {ISO timestamp}
modified: {ISO timestamp}
---

evergreen and terms tags are required — they're how the vault identifies these notes.

Body shape — encountered mode:

[[Source note]]:
> Verbatim quote that defines or illustrates the term. Multiple paragraphs of the source can be stacked as separate blockquote lines.
> Another excerpt if relevant.

## See also
- [[Related term]]
- [[Related concept]]

Body shape — coined mode:

1–2 sentence gloss in plain language. The phrase or context that prompted it can sit below as a usage example or as a quoted line.

> Optional: original phrase being replaced, or a sentence using the term.

## See also
- [[Related term]]

6. Confirm

Show the resulting wikilink ([[Term name]]) and a one-line summary of what was saved. If a daily-log bullet is warranted (per the vault's write-back discipline), mention it but don't auto-write — that's a separate decision.

Tool choice

Per AGENTS.md, prefer Obsidian CLI when the desktop app is running. For new file creation in a known location, the Write tool is fine and direct. Read existing Term files with obsidian read file="<term>" to check for collisions and to study tone/length when in doubt.

Examples

Encountered:

User: /save-term affordance
[skill asks for source]
User: From Don Norman's Design of Everyday Things — there's a clipping at [[The Design of Everyday Things - Norman]]
[skill drafts the note with that source + a quote pulled from the clipping]

Coined (post phrase-revision):

User: /save-term ambient drafting
[skill confirms it's coined, asks for a quick gloss]
User: writing in the background while doing other things — the prose accretes rather than gets composed
[skill drafts a short coined-mode note, suggests `LiPenLine` for icon]

What not to do

  • Don't paraphrase source quotes in encountered mode. Verbatim only.
  • Don't invent metadata fields not in the existing convention (no category, no type, etc.) — keep frontmatter minimal.
  • Don't write ambient prose around the term. Term notes are tight scrapbook entries, not essays.
  • Don't set publish: true on a new Term note. Default is publish: false so the note can ripen privately first; the user flips it on themselves when it's ready.
  • Don't auto-update STATUS or other vault state. This skill writes one file.