agentleFS
fahdAghribifahdAghribi/youtube-channel-skills/skills/youtube-publisher/SKILL.md

youtube-publisher

Skill2 starsChanged 14 days agoMITView on GitHub

Upload, schedule and update YouTube videos through the official Data API with a fingerprinted approval gate: dry run shows the exact payload, the user approves, only then it publishes. Handles thumbnail, playlist, scheduling, visibility changes and metadata edits on existing videos. Not for writing metadata (use youtube-metadata-optimizer).

4 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 0 are distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

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

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

    Make sure a video.json exists and passes python3 scripts/validate_metadata.py video.json.

    line 27
  2. Very specific:

    Write changes.json with only the fields to change (title, description, tags, privacyStatus, publishAt, madeForKids, thumbnail).

    line 37
  3. Very specific:

    If the Google Cloud project is unaudited, uploads stay private regardless (see ../../references/api-quota.md gotcha 1).

    line 46
  4. Very specific:

    Never change madeForKids or make a video public as a side effect of another request.

    line 52

“Common” is measured: kept by at least 6 in 10 of the 4 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

YouTube Publisher

The only skill that writes to the channel's videos. Every action is two-pass: a dry run that prints an approval card with a fingerprint of the exact payload, and a confirmed run with --confirm <fingerprint> that refuses to act if anything changed.

When to use

  • "Upload this video", "schedule it for Tuesday 18:00", "make it public"
  • "Change the title/description/thumbnail of this video"
  • Bulk updates (e.g. adding a link to the last 20 descriptions), one approval per video

Prerequisites

Tier 2 (owner). Check with python3 scripts/check_config.py. If not ready, walk the user through README.md §Setup (Google Cloud project, OAuth Desktop client, python3 scripts/auth.py). Without Tier 2, fall back to manual mode: give a Studio checklist with every field ready to paste.

Steps: new upload

  1. Make sure a video.json exists and passes python3 scripts/validate_metadata.py video.json. If not, run youtube-metadata-optimizer first.
  2. Confirm the channel: check_config.py / auth.py printed which channel the token controls. Say it to the user before the first write in the session.
  3. Dry run: python3 scripts/publish_video.py video.json. Show the approval card to the user verbatim (title, description, fields, validation, fingerprint).
  4. Wait for explicit approval ("yes", "sí", "publica", "dale"). Edits → update video.json → new dry run → new fingerprint.
  5. Confirmed run: python3 scripts/publish_video.py video.json --confirm <fingerprint from the card>.
  6. Report the video URL, visibility, scheduled time (in the user's time zone), thumbnail/playlist results.
  7. Studio-only handoff: pinned comment, end screens, cards, Test & Compare variants, subtitles check. Give the checklist from the metadata step.

Steps: update an existing video

  1. Write changes.json with only the fields to change (title, description, tags, privacyStatus, publishAt, madeForKids, thumbnail).
  2. Dry run: python3 scripts/update_video.py <url> changes.json shows a before/after diff and fingerprint.
  3. Approval → python3 scripts/update_video.py <url> changes.json --confirm <fingerprint>.

Scheduling rules

  • publishAt is UTC ISO 8601 in the file; always tell the user the local time too.
  • Scheduled uploads must be private until publish time (validator enforces it).
  • Upload ≥ 24 h ahead so HD processing and checks finish.
  • If the Google Cloud project is unaudited, uploads stay private regardless (see ../../references/api-quota.md gotcha 1). Tell the user before uploading, and offer to schedule from Studio instead.

Hard rules

  • Never run with --confirm without the user's explicit approval of that specific card in this conversation. Never compute or guess a fingerprint yourself; copy it from the card the user saw.
  • Approval for one video does not carry to the next one, nor to an edited payload.
  • Never change madeForKids or make a video public as a side effect of another request.
  • Bulk operations: one card per video, or one card listing every video with its own fingerprint; the user approves each.
  • On failure, report the error as returned. Do not retry uploads blindly (each attempt costs quota).

Files

  • scripts/publish_video.py, scripts/update_video.py, scripts/validate_metadata.py, scripts/check_config.py, scripts/auth.py
  • ../../references/api-quota.md
  • ../../references/metadata-limits.md

Related skills

  • youtube-metadata-optimizer — produces the video.json
  • youtube-title-thumbnail — the thumbnail file and title
  • youtube-comment-manager — first-hours comment replies after publishing

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.