Skip to content
SkillMixed3,059 tokens in use

Packaging YouTube Thumbnails

Creates scored YouTube title and thumbnail packages from a saved channel profile and source pack, renders and checks them, and learns from Studio metrics.

3,862 stars on the repository MIT Licenseericosiu/ai-marketing-skills · packaging-youtube-thumbnails
npx skills add ericosiu/ai-marketing-skills -s packaging-youtube-thumbnails
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing Packaging YouTube Thumbnails costs

30tokens idle
3,059tokens in use
7,197tokens, up to

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Up to adds the files its instructions link. Counts are approximate, made offline with cl100k_base.

Model dependence Mixed

Script-ledMixedModel-led

Scripts do part of the work and the model's judgement does the rest.

This describes the skill's structure, how much of it is scripts. It does not say how well it works. Scripts: scripts/thumbnail_guard.py, scripts/thumbnail_learning.py.

Touches

shelldetectedwrites filesdetected

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Excellent
  • Frontmatter8/15
  • Structure25/25
  • Examples12/15
  • Output15/15
  • Size15/15
  • Resources6/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent
  • Runs shell commands−5

    does this on the machine, and the frontmatter does not say so

    SKILL.md, line 19
  • Writes files−8

    does this on the machine, and the frontmatter does not say so

    scripts/thumbnail_learning.py, line 62

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity88
  • Maintenance44
  • Community56
  • Maturity28
  • Momentum40

Packaging YouTube Thumbnails turns a transcript, recording or brief into differentiated title and thumbnail packages. It loads a saved channel profile, reads any performance ledger, writes packages on three distinct promise lanes, scores them against a rubric and recommends one.

Renders use an image-generation skill when one is available, then pass numeric safe-zone and edit-region checks from a bundled Python guard. After publishing, the skill stores Studio snapshots and runs a postmortem.

What You Get

  • A validated channel profile with identity rules, approved references and an output folder
  • The requested number of packages, three by default, each with exact title, thumbnail copy, composition, hook logic and risk
  • Scores for every package and one recommendation
  • Rendered variants saved to non-overwriting versioned paths, shown inline with numeric QA evidence
  • A tested ZIP handoff for an editor and a local ledger with 24-hour, 72-hour and seven-day snapshots

Common Use Cases

  • Packaging a new episode with three different promise hypotheses
  • Revising one thumbnail while leaving approved variants untouched
  • Reviewing how a published package performed against comparable ones

What Makes It Unique

The skill treats each package as a testable promise and grades evidence by class: a direct user correction applies at once, public views are only directional, and a lesson needs three distinct 72-hour readings and explicit review before it is saved. It translates words like wins or easier into object size and position, runs a copy-off test, and guards surgical edits by diffing the pixels outside the allowed region.

Under The Hood

The skill holds instructions, four reference files and two Python scripts that need the Pillow imaging library. The guard script validates the profile file, image dimensions, safe margins and the unchanged region of an edited image. The learning script reads and writes a JSON lines ledger on disk with a file lock. At the start it runs the repository’s version check and telemetry helper through the shell, with remote telemetry opt-in. The agent is told to read source material, inspect recent thumbnails of the channel for a first profile, use an image-generation skill, resolve brand marks from local files or official pages, and write images, a manifest and a ZIP. It does not publish or change a live YouTube asset without approval.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers