agentleFS
eugenezhangco-speceugenezhangco-spec/athena/.claude/skills/transcribe/SKILL.md

transcribe

Skill0 starsChanged 7 months agoMITView on GitHub

Drop a video or audio URL — Athena fetches the audio, transcribes it, and returns a clean transcript plus a short summary. Works on YouTube, Instagram reels, TikTok, podcasts, and any yt-dlp-supported platform. Use when the user shares a link and wants the content extracted, not just bookmarked.

  • Reads credentials
  • Deletes or force-pushes
  • Installs packages

51 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use spreadsheet, think and 9 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • JetBrainsSkill

    spreadsheet

    Use when tasks involve creating, editing, analyzing, or formatting spreadsheets (`.xlsx`, `.csv`, `.tsv`) with formula-aware workflows, cached recalculation, and visual review.

    Also in 3 projects that use this

  • vltingSkill

    think

    Lightweight plan-to-execution pipeline. Explores codebase, runs council deliberation, presents plan for approval, then executes via /run.

    Also in 3 projects that use this

  • robzolkosSkill

    done

    Reset the working directory for the next task by ensuring no uncommitted changes exist, then switching to the main branch and pulling latest.

    Also in 3 projects that use this

  • JaimeGBASkill

    ingest

    Digest an external source (article, paper, transcript, URL, book chapter) and integrate its content into the knowledge base

    Also in 3 projects that use this

  • AnthropicSkill

    onboarding

    Generate an onboarding checklist and first-week plan for a new hire. Use when someone has a start date coming up, building the pre-start task list (accounts, equipment, buddy), scheduling Day 1 and Week 1, or setting 30/60/90-day goals for a new team member.

    Also in 3 projects that use this

  • MicrosoftSkill

    diagnose

    Use to systematically diagnose why scenarios fail — identifies root causes by analyzing agent traces, codebase, and behavioral patterns

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

10 rules. 0 are common, 9 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 44 edited versions we read keep these.

  • 9

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If neither Groq nor whisper-cli is available, tell the user how to set up Groq (free tier, 30s) and stop.

    line 47
  2. Distinctive:

    If user dropped a URL with no other instruction, ask one short question: "Want me to transcribe it?" Wait for confirmation.

    line 55
  3. Distinctive:

    If the transcript is short (under 500 words), inline the full text.

    line 107
  4. Very specific:

    Don't editorialize the transcript itself — it's the speaker's words, not Athena's interpretation.

    line 108
  5. Distinctive:

    If the audio is mostly music or non-verbal, say so: "Mostly background music, no transcribable speech." Don't fabricate.

    line 110
  6. Distinctive:

    Does NOT chain to content-creation skills (no auto-remix to LinkedIn posts, etc.).

    line 116
  7. Distinctive:

    Does NOT save to long-term memory automatically — user must ask.

    line 117
  8. Distinctive:

    Does NOT translate. If the transcript comes back in another language, deliver it as-is and offer translation as a follow-up.

    line 118
  9. Distinctive:

    Does NOT extract video frames or screenshots — audio only.

    line 119
  10. Distinctive:

    Does NOT work on protected content where yt-dlp can't access (DRM, paywall).

    line 120

“Common” is measured: kept by at least 6 in 10 of the 44 edited versions of this file we could read. “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

Transcribe

Pulls audio from any supported URL, transcribes it, and surfaces the content. Pure ingestion — what the user does with it (memory, notes, content remix) is downstream.


When to Trigger

Auto-fire on URL drops — if the user pastes a link from any of these platforms with no other instruction, ask "Want me to transcribe it?" then run:

PlatformMethod
YouTube / Shortsyt-dlp → audio → transcribe
Instagram Reelsyt-dlp (with cookies) → audio → transcribe
TikTokyt-dlp → audio → transcribe
Podcasts (RSS/MP3 link)direct download → transcribe
Twitter/X videoyt-dlp → audio → transcribe
Any audio/video URL yt-dlp acceptssame flow

Explicit triggers:

  • "Transcribe this"
  • "What does this video say"
  • "Summarize this clip"
  • "Pull the transcript from [URL]"
  • "/transcribe [URL]"

Do NOT trigger for:

  • Article links (use WebFetch instead — no audio to transcribe)
  • Links the user has clearly already watched and is referencing
  • The user pasting a link to share with Athena, not to ingest

Requirements

ToolInstallPurpose
yt-dlpMac: brew install yt-dlp / Ubuntu: pipx install yt-dlpFetches audio from any video URL
ffmpegMac: brew install ffmpeg / Ubuntu: apt-get install ffmpegAudio conversion
GROQ_API_KEYFree tier at console.groq.com — set in bot/.env or .envWhisper transcription (primary path)
whisper-cli (optional, Mac fallback)brew install whisper-cppLocal transcription if no Groq key

If neither Groq nor whisper-cli is available, tell the user how to set up Groq (free tier, 30s) and stop.


Steps

Step 1: Confirm intent

If user dropped a URL with no other instruction, ask one short question: "Want me to transcribe it?" Wait for confirmation. Skip this if user explicitly said transcribe.

Step 2: Fetch audio

bash scripts/fetch-audio.sh "<URL>" /tmp/transcribe-$$

Returns the path to audio.mp3. If yt-dlp fails (private video, geo-block, expired link), report cleanly: "Couldn't fetch — [reason]. Got a different link?"

Step 3: Transcribe

bash scripts/transcribe.sh /tmp/transcribe-$$/audio.mp3

Returns plain text on stdout. Capture it.

Step 4: Clean up

rm -rf /tmp/transcribe-$$

Step 5: Deliver

Output format:

**Transcript** ([duration] · [platform])

[3-5 sentence summary — what the speaker actually says, not metadata. Lead with the thesis.]

---

**Full transcript:**
[Full text. If over ~3000 words, save to `personal/inbox/transcripts/<slug>-<date>.md` and link instead.]

Step 6: Offer next moves (one line, not a menu)

Pick the most likely follow-up based on context:

  • Coaching/learning content → "Want me to extract the key takeaways?"
  • Long-form interview/podcast → "Want me to save this to your inbox?"
  • Tutorial/how-to → "Want me to pull out the action items?"
  • Otherwise → no follow-up offer; just end clean.

Output Rules

  • Lead with the summary, not the metadata. The user wants to know what was said.
  • If the transcript is short (under 500 words), inline the full text. If long, save to disk and link.
  • Don't editorialize the transcript itself — it's the speaker's words, not Athena's interpretation.
  • The summary IS Athena's interpretation — be sharp, not neutral.
  • If the audio is mostly music or non-verbal, say so: "Mostly background music, no transcribable speech." Don't fabricate.

What This Skill Does NOT Do

  • Does NOT chain to content-creation skills (no auto-remix to LinkedIn posts, etc.).
  • Does NOT save to long-term memory automatically — user must ask.
  • Does NOT translate. If the transcript comes back in another language, deliver it as-is and offer translation as a follow-up.
  • Does NOT extract video frames or screenshots — audio only.
  • Does NOT work on protected content where yt-dlp can't access (DRM, paywall). Report and stop.

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.