agentsclimarketplace

Kol quote research

Skill Ryder-MHumble/Youtube-KOL-Transcripts/skills/kol-quote-research

Stop replaying hours of YouTube interviews. Get continuously updated, timestamped KOL transcripts, exact quote retrieval, and a portable Obsidian knowledge base for strategy, investment, product, and AI research.

Install
npx -y skills add Ryder-MHumble/Youtube-KOL-Transcripts --skill kol-quote-research

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

3 things to look at

  • 13 days oldThe repository was created 13 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Retrieve evidence-backed viewpoints from the kol-transcripts repository. Use when a user wants to compare a thesis with KOL opinions, find who expressed a similar or opposing view, recover exact original wording from transcript Markdown files, or cite a video URL and timestamp. Especially useful for strategy research, market analysis, literature reviews, and claim verification across English or Chinese queries.

SKILL.md

2.7 KB, 456 tokens by cl100k_base, as published. Nobody here has run it

KOL Quote Research

Find relevant transcript passages, verify them in context, and return exact quotations with traceable sources.

Workflow

  1. Locate the repository root containing catalog.jsonl and transcripts/.
  2. Rewrite the user's claim as a neutral research question.
  3. Expand the query into both Chinese and English search terms. Add synonyms, named concepts, and one opposing formulation. English expansion is required when the transcripts are English.
  4. Run the bundled search script two or three times with different formulations:
python3 skills/kol-quote-research/scripts/search_transcripts.py \
  "implementation cost execution becomes cheap scarce judgment taste strategy" \
  --limit 12 --json
  1. Open the highest-scoring transcript files and read the quoted passage plus adjacent timestamps. Search results are candidates, not final evidence.
  2. Keep only passages that directly support, qualify, or contradict the user's claim.
  3. Preserve the exact source wording. Put translations or summaries in a separate field.

Attribution Rules

  • Treat speaker_attribution: contextual as an explicit warning: the transcript has timestamps but no diarized speaker labels.
  • Name a person only when the title, nearby dialogue, or metadata makes the speaker clear.
  • Otherwise attribute the quote to the publishing account and label the speaker as unverified.
  • Never turn a channel description, chapter title, interviewer question, or paraphrase into a KOL quote.
  • Never invent a timestamp. If no timestamp exists, state that limitation.

Required Output

Return a compact evidence table with these fields:

  • Relationship to the user's claim: aligned, qualified, or opposed
  • KOL or account
  • Exact original quote
  • Optional Chinese translation when the quote is not Chinese
  • Source video and clickable timestamp
  • One-sentence context
  • Attribution confidence: high, medium, or unverified

After the table, synthesize agreements and disagreements. Separate evidence from interpretation. If no direct evidence exists, say so instead of filling the gap with general knowledge.

Resources

  • Use scripts/search_transcripts.py for ranked passage retrieval.
  • Read references/schema.md when metadata, links, or attribution fields are unclear.

Gives 0 of the 12 instructions most research analysis skills give in 456 tokens

Counted across 1,063 of the 1,754 authors here whose files we hold, read 2026-08-06

  • generate a markdown reportin 32 of 1063, across 17 files
  • cite each claim's sourcein 31 of 1063, across 14 files
  • define the ideal customer profilein 20 of 1063, across 2 files
  • search for companies matching the criteriain 20 of 1063, across 2 files
  • assign a fit score from one to tenin 20 of 1063, across 2 files
  • format results in a scannable markdown templatein 20 of 1063, across 2 files
  • analyze the codebase to understand the productin 19 of 1063, across 1 file
  • ask clarifying questions about the value propositionin 19 of 1063, across 1 file
  • look for signals of immediate needin 19 of 1063, across 1 file
  • identify the target decision maker rolein 19 of 1063, across 1 file
  • suggest a personalized contact strategyin 19 of 1063, across 1 file
  • provide conversation starters for outreachin 19 of 1063, across 1 file

Said here and by no other author read

  • rewrite the user claim as a neutral research question
  • expand the query into both English and Chinese search terms
  • add synonyms named concepts and one opposing formulation
  • run the search script two or three times
  • read the quoted passage plus adjacent timestamps
  • keep only passages that support qualify or contradict

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.