How do I install the Research To Vault skill?

Run npx skillmds@latest add adamradek-ux/research-to-vault in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Research To Vault safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Research To Vault?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Research To Vault free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Research To Vault?

adamradek-ux (@adamradek-ux) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Research To Vault by adamradek-ux — Agent Skill

Name: Research To Vault
Author: adamradek-ux

Research To Vault

Takes a URL, YouTube video, PDF/DOCX, or a voice memo, extracts clean content, writes a short summary, and PROPOSES where to file it in your notes vault (resources/ or inbox/) with proper frontmatter and links. Propose → OK → write. Triggers: "save this article/video", "process this source into my vault", "summarize and file", "research to vault", a pasted URL/file to save, and Czech aliases "ulož tenhle článek", "shrň a zařaď".

adamradek-ux Updated Sep 17, 2026

File contents

Turn an external source (article, video, PDF, voice memo) into a clean vault note: extract → summarize → propose filing. Do not write without an OK.

Procedure

1. Extract clean content (reuse tools, don't reimplement)

By source type:

Web article / blog → a clean-extraction skill if you have one (e.g. an Obsidian/defuddle-style markdown extractor). Fallback: WebFetch. For .md URLs use WebFetch directly.
YouTube / PDF / DOCX / PPTX / image → a markdown converter if available (e.g. markitdown: YouTube yields the transcript, PDF/DOCX the text).
Audio / voice memo → a transcription tool (e.g. a Whisper-based CLI).

If extraction fails, say so — do not invent content.

2. Summary (not a transcript)

3–6 bullet points of what matters most + one sentence on "why this is relevant to the user".
Keep key numbers/names/URLs verbatim; mark author claims as claims.
Cite the source (URL/channel/author) — every fact must be traceable.

3. Propose filing (GTD-style triage)

Criterion	Target
Reference/guide/material (not actionable)	`resources/` (or your reference folder)
Deadline + multiple steps	your projects folder (offer it; rare)
Unclear / to be sorted	`inbox/`

Propose a concrete path + file name (kebab-case, no diacritics) and the note body:

---
tags: [domain, subtype]
status: active
created: YYYY-MM-DD
source: <URL or origin>
---

# <Title>

<summary>

## Source
<URL / channel / author>

If your vault uses [[wikilinks]], add 1–2 links to existing related notes — find them, don't guess them.

4. Propose → OK → write

List the target path + content, wait for the OK. After writing you can suggest verify-before-done.

Conventions

Kebab-case without diacritics, no placeholders, absolute dates.
Frontmatter per your vault's convention (e.g. tags/status/created).
Prefer a tight summary over a long transcript — the vault is for future you, not an archive of raw text.

---name: research-to-vaultdescription: Takes a URL, YouTube video, PDF/DOCX, or a voice memo, extracts clean content, writes a short summary, and PROPOSES where to file it in your notes vault (resources/ or inbox/) with proper frontmatter and links. Propose → OK → write. Triggers: "save this article/video", "process this source into my vault", "summarize and file", "research to vault", a pasted URL/file to save, and Czech aliases "ulož tenhle článek", "shrň a zařaď".---Turn an external source (article, video, PDF, voice memo) into a clean vault note:extract → summarize → **propose filing**. Do not write without an OK.## Procedure### 1. Extract clean content (reuse tools, don't reimplement)By source type:- **Web article / blog** → a clean-extraction skill if you have one (e.g. an Obsidian/defuddle-style markdown extractor). Fallback: WebFetch. For `.md` URLs use WebFetch directly.- **YouTube / PDF / DOCX / PPTX / image** → a markdown converter if available (e.g. [markitdown](https://github.com/microsoft/markitdown): YouTube yields the transcript, PDF/DOCX the text).- **Audio / voice memo** → a transcription tool (e.g. a Whisper-based CLI).If extraction fails, say so — do not invent content.### 2. Summary (not a transcript)- 3–6 bullet points of what matters most + one sentence on "why this is relevant to the user".- Keep key numbers/names/URLs verbatim; mark author claims as claims.- Cite the source (URL/channel/author) — every fact must be traceable.### 3. Propose filing (GTD-style triage)| Criterion | Target ||---|---|| Reference/guide/material (not actionable) | `resources/` (or your reference folder) || Deadline + multiple steps | your projects folder (offer it; rare) || Unclear / to be sorted | `inbox/` |Propose a concrete **path + file name** (kebab-case, no diacritics) and the note body:```---tags: [domain, subtype]status: activecreated: YYYY-MM-DDsource: <URL or origin>---# <Title><summary>## Source<URL / channel / author>```If your vault uses `[[wikilinks]]`, add 1–2 links to existing related notes — findthem, don't guess them.### 4. Propose → OK → writeList the target path + content, wait for the OK. After writing you can suggest`verify-before-done`.## Conventions- Kebab-case without diacritics, no placeholders, absolute dates.- Frontmatter per your vault's convention (e.g. `tags/status/created`).- Prefer a tight summary over a long transcript — the vault is for future you, not an archive of raw text.

Frequently asked questions