agentleFS
Sign inStart free
utillayerutillayer/youtube-agent-skill/skills/yt-research/SKILL.md

yt-research

Skill10 starsChanged 29 days agoMITView on GitHub

Research a video from its primary source before anything is written: a cited brief of what the source actually says, then an independent fact check of the finished script. Use for history, mythology, scripture, science or any video that retells a text or a record - "research this", "what does the original say", "fact check my script", "is this accurate", "find the source".

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. qwwiwi/agentos-skills-public★ 4same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository11 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If the user cannot share or name a source, stop and say the video cannot be checked.

    line 35

≈ “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

yt-research

A retelling channel lives or dies on one promise: what you say happened is what the source says happened. This skill keeps it. It works in two passes, and the second is done by someone who did not do the first.

python3 ${CLAUDE_SKILL_DIR}/sourcekit.py index source.txt          # sections with line ranges
python3 ${CLAUDE_SKILL_DIR}/sourcekit.py find  source.txt "phrase"   # where the source says it
python3 ${CLAUDE_SKILL_DIR}/sourcekit.py show  source.txt "Canto XLV"

The tools live in this skill's folder. ${CLAUDE_SKILL_DIR} is that folder in Claude Code; anywhere it does not expand (claude.ai, Codex), use the folder this SKILL.md was loaded from.

1. Get the primary source

The text itself, not a summary of it and not a video about it. Prefer a public-domain translation or the original record: Project Gutenberg, sacred-texts.com, archive.org, official transcripts. Save it as plain text in the project and write down what it is: translator, year, edition, what it leaves out. If the edition numbers its chapters differently from other editions, say so now, and cite in this edition's numbering every time.

If the user cannot share or name a source, stop and say the video cannot be checked. Do not fall back on memory and present it as the text.

2. The brief (brief.md)

Index the source, read the sections the video covers - in full, footnotes included - and write:

  • Events in order, each with its section and line range: "Book III, Canto XLV, lines 32700-32790".
  • Words worth quoting, short, with line numbers.
  • The careful list: what is ambiguous, what the translator flags as a later insertion, what the source omits, and every famous detail that comes from a later retelling rather than this text. The script may mention those, but only as what they are ("later retellings add...").

The brief is what the script is written from. If it is not in the brief, it is not in the script.

3. The independent fact check (after the script is written)

Whoever wrote the script does not check it. Start a fresh agent that never saw the writing and give it this, filled in:

You are fact-checking a narration script against its source. You did not write it. Script: <path>. Source: <path>, lines <a>-<b> (<edition>, <sections>). Read those lines in full, footnotes included. For every factual claim (who did what, numbers, names, order of events, speech), return one row: | Para | Claim | Verdict | Evidence (line numbers, short quote) | Verdicts: SUPPORTED (the source says it) · NOT IN SOURCE (invention or later tradition; say which if you know) · CONTRADICTED (the source says otherwise) · OVERSTATED (true in spirit, stronger or more specific than the text) · FLAGGED (the source has it, but the editor calls the passage spurious). No line number, no SUPPORTED. Paraphrase is fine; changed meaning is not. Also list anything the script leaves out that would mislead a viewer. End with the count of each verdict. Write the table to factcheck.md. Do not edit the script.

Gate: nothing is recorded until factcheck.md has no CONTRADICTED, and every NOT IN SOURCE, OVERSTATED and FLAGGED claim is fixed or said plainly in the narration.

What to hand back

The brief, the source's line ranges, and the careful list. After the check: the verdict counts and each claim that needs fixing, with the line that settles it.

Before you hand it over

This skill drafts; you publish. End with the copy-ready block and one question: post it as is, or another pass?

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.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

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