Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add fredoliveira/agents --skill youtube-to-atomsgit clone --depth 1 https://github.com/fredoliveira/agentsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/fredoliveira/agents/youtube-to-atoms)<a href="https://agentmods.dev/skills/fredoliveira/agents/youtube-to-atoms"><img src="https://agentmods.dev/badge/skills/fredoliveira/agents/youtube-to-atoms/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/fredoliveira/agents/youtube-to-atoms"><img src="https://agentmods.dev/badge/skills/fredoliveira/agents/youtube-to-atoms.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00050 | $0.01038 |
| Opus 5 | $0.00025 | $0.00519 |
| Sonnet 5 | $0.00010 | $0.00208 |
| Haiku 4.5 | $0.00005 | $0.00104 |
Grade A, and why
youtube-to-atoms scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 104 lines — stays where its author put it; the contents beside it link to each section on GitHub.
YouTube → atomic notes
Purpose
Turn one YouTube video into a source note plus a set of atomic notes in an Obsidian vault. The unit of extraction is a solid, transferable idea or concept — not a summary.
The note style (what an atomic note is, frontmatter, linking, a worked example) is fully codified in the bundled reference.md — this skill is self-contained and does not depend on any particular vault existing beforehand.
Configuration
The vault root is the YT_VAULT environment variable, defaulting to
~/Documents/obsidian/yt. The skill reads/writes three subfolders (created
automatically):
$YT_VAULT/Transcripts/— fetched transcript (provenance)$YT_VAULT/Sources/— one source note per video$YT_VAULT/Atoms/— the atomic notes
To target a different vault, set YT_VAULT before invoking (e.g.
export YT_VAULT=~/notes/myvault).
Instructions
1. Fetch the transcript
Run the bundled script with the URL the user gave you:
python3 "${CLAUDE_SKILL_DIR}/fetch_transcript.py" "<youtube url>"
It prints the absolute path of the transcript markdown file (or an ERROR: line
if the video has no English captions, or if YouTube throttled the request —
relay that to the user and stop). If a transcript for that video already exists,
it prints the existing path.
Requires
yt-dlp(python3 -m pip install yt-dlp). Auto-generated captions are used when no human captions exist — they extract fine.
2. Read the transcript IN FULL
Read the printed file completely (page through with offset/limit if it's long —
many videos run for hours). Do not extract from a partial read. The
frontmatter holds url, channel, duration; body paragraphs are prefixed
with **[MM:SS]** timestamps.
3. Learn the style, then extract
Read reference.md (in this skill's directory) — it is the authoritative guide for note shape, linking, and the worked example. Then extract atomic notes from the transcript following it. The essentials:
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 104 lines · 50 tokens per session scan A 824cbc459cd5
youtube-to-atoms is a skill published in the GitHub repository fredoliveira/agents (24 stars, last pushed 2mo ago), licensed MIT. It adds 50 tokens to every session and 1,038 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
learn
Socratic teaching for explicit learning requests: diagnose prior understanding, teach one concept at a time, verify comprehension, and correct misconceptions. Use for /learn, "teach me", or "I don't get X".
tutorial
Checkpointed tutorial writer for step-by-step technical teaching: visible end state, dependency-ordered steps, working checkpoints, and mechanism explanations. Use when the user asks to create, draft, revise, or outline a tutorial, step-by-step guide, onboarding guide, or documentation that teaches a skill.
video-tutorial-maker
Create scripted tutorial videos with narration, TTS, aligned subtitles, MP4 rendering, and platform variants. Use when Codex needs to create or revise tutorial, explainer, product walkthrough, course, demo, or short-form videos; generate 16:9 landscape output by default unless the user explicitly asks for 9:16, 1:1…
openehr-assistant
This skill should be used when a conversation touches openEHR outside a task owned by a dedicated skill — e.g. "what is an archetype?", "how do openEHR templates work?", "which composition category fits this data?", "find me a guide on X", "where do I start with openEHR modeling?" — or names openEHR concepts (ADL…
dictionary-get-{word}
Returns the definition, phonetics, examples, and synonyms for the given word.
dictionary-post-batch
Resolve multiple words in a single request. Returns dictionary entries when found, or error information when a word is not in the dataset.