agentleFS
Sign inStart free
Murtadha-NajemMurtadha-Najem/claude-reel/SKILL.md

reel

Skill0 starsChanged 55 days agoMITView on GitHub

Understand an Instagram reel, video post or photo post from its link. Use whenever the user sends an instagram.com/reel, /reels, /p or /tv link, or asks what a reel says or shows, even with no instruction beyond the link ("what's in this reel", "summarise this", "شنو بهذا الريل", "/reel <url>"). Downloads it, reads caption, metadata, audio and video with cheap local checks first, and spends Gemini or image tokens only where they are needed.

10 other projects use a version of this, 1 of them edited.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 1 is distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 7 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use the Python environment the requirements were installed into.

    line 12
  2. Very specific:

    If it prints REEL ERROR, relay the message plainly.

    line 16
  3. Very specific:

    Read bundle.md in full. Text comes first: caption, Instagram metadata, song, speech, on-screen text.

    line 19
  4. Very specific:

    If something is still unclear, look further with the look.py commands printed at the end of the bundle: a sheet of any time range, or one …

    line 21
  5. Distinctive:

    Never reproduce full song lyrics.

    line 30
  6. Very specific:

    Write the file named on the record file to write line:

    line 33

“Common” is measured: kept by at least 6 in 10 of the 7 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

/reel

Run

python ~/.claude/skills/reel/reel.py "<url>"

Use the Python environment the requirements were installed into. The script prints the path of bundle.md.

Flags: --force-transcribe when the on-screen captions look incomplete, --refresh to download again, --max-frames N to change the frame cap (default 8), --dense when something shown within one camera shot matters (a product reveal, a prop, a gesture).

If it prints REEL ERROR, relay the message plainly. A cookies error means the user must export Instagram cookies (a browser extension such as "Get cookies.txt LOCALLY", while logged in) to ~/.config/reel/cookies.txt. Never ask them to paste cookie contents into chat.

Read

  1. Read bundle.md in full. Text comes first: caption, Instagram metadata, song, speech, on-screen text.
  2. For a video, open the overview sheet it names. Each cell is a detected moment with its time.
  3. If something is still unclear, look further with the look.py commands printed at the end of the bundle: a sheet of any time range, or one full frame to read small text. Stop as soon as the post is understood.
  4. For a photo post, read the frames listed under "Frames to read".
  5. The Decisions section says what was skipped and why. If a skip looks wrong for this post (captions clearly do not match what is said), rerun with the relevant flag rather than guessing.

Answer

Reply in the user's language. Say what the post is (format, who, topic), what is said, what is shown, the song if any, and whether the caption relates to the content or is just bait. Keep it proportional: a 10 second meme needs two lines.

Do not invent speech that is not in the bundle. If the speech source says "NOT transcribed", the post has speech nobody transcribed (no Gemini key, or the daily free quota is spent): say so, describe only what the frames and on-screen text show, and offer to rerun later. If the speech source is "none" and the audio kind is music, the post has no spoken content.

Never reproduce full song lyrics; name the song and describe it.

Record

Write the file named on the record file to write line:

---
url: <url>
account: <@account>
posted: <date>
song: <title by artist, or none>
---
# <one-line description>

<summary>

## Speech
<transcript or captions text, or none>

## On-screen text
<lines, or none>

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