Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add alebgl77/claude-inc --skill youtube-thumbnailgit clone --depth 1 https://github.com/alebgl77/claude-incWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/alebgl77/claude-inc/youtube-thumbnail)<a href="https://agentmods.dev/skills/alebgl77/claude-inc/youtube-thumbnail"><img src="https://agentmods.dev/badge/skills/alebgl77/claude-inc/youtube-thumbnail.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00120 | $0.01033 |
| Opus 5 | $0.00060 | $0.00517 |
| Sonnet 5 | $0.00024 | $0.00207 |
| Haiku 4.5 | $0.00012 | $0.00103 |
Grade A, and why
youtube-thumbnail scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 86 lines — stays where its author put it; the contents beside it link to each section on GitHub.
YouTube Thumbnail — Cover Tester
"Test thumbnail covers"
When to use
- "Thumbnail ideas for my video about ."
- "Which of these two covers should I use?" — a scored verdict, not a shrug.
- CTR is flat and packaging is the suspect: "my CTR is stuck at 3%."
- Pre-production: designing title + thumbnail as one package before filming.
Workflow
- Get the inputs: video topic, the exact title, the target viewer, and where the impressions come from (browse, suggested, search) — thumbnails compete differently on each surface.
- Define the ONE emotion or question the thumbnail must trigger. The thumbnail asks a question the title answers (or the reverse) — the two must never repeat each other.
- Draft 3–5 concept briefs. Each specifies composition (focal point on a rule-of-thirds intersection), face or object (expression, prop, angle), a text overlay of three words or fewer (or none), and a palette with deliberate contrast against YouTube's white/dark UI and red accents.
- Curiosity-check each concept: does it open a loop the title closes? Does it spoil the payoff? Does it still make sense with the title covered?
- Run the mobile legibility test: judge each concept at 168×94 px — focal point readable, overlay text at 25%+ of frame height, nothing critical under the timestamp corner or lost at the edges.
- Score every concept /10 with the rubric: clarity at a glance (3), curiosity (3), contrast vs UI (2), title synergy (2). Rank them.
- If an image generation tool is available in the session, produce mockups of the top two concepts; otherwise ship briefs precise enough for a designer or a Canva session.
- Write the A/B rotation plan: which concept launches, the numeric swap trigger (e.g. CTR under 4% after 48 hours or 10k impressions), and what the variant changes — exactly one variable.
Output format
## Thumbnails: "<video title>"
Target emotion/question: <one line> | Surface: browse / suggested / search
### Concept 1 — <name>
- Composition: <focal point + layout>
- Face/object: <expression or prop>
- Overlay: "<3 words max>" | none
- Colours: <palette + contrast note vs YouTube UI>
- Curiosity check: pass / fail — <reason>
- Mobile check (168×94): pass / fail — <reason>
- Score: <n>/10 (clarity /3, curiosity /3, contrast /2, synergy /2)
### Concept 2–5 — <same structure>
### Verdict
Launch: Concept <n> — <one-line reason>
A/B plan: swap to Concept <m> if <numeric trigger>; variable changed: <one thing>
Mockups: <file paths | "no image tool available — briefs are designer-ready">
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 86 lines · 120 tokens per session scan A 5d802bc426c9
youtube-thumbnail is a skill published in the GitHub repository alebgl77/claude-inc (14 stars, last pushed 5d ago), licensed MIT. It adds 120 tokens to every session and 1,033 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-04.
Other skills, from other repositories
edge-tts
Text-to-speech conversion using uvx edge-tts for generating audio from text. Use when (1) User requests audio/voice output with the "tts" trigger or keyword. (2) Content needs to be spoken rather than read (multitasking, accessibility, driving, cooking). (3) User wants a specific voice, speed, pitch, or format for TTS…
zai-tts
Text-to-speech conversion using GLM-TTS service via the uvx zai-tts command for generating audio from text. Use when (1) User requests audio/voice output with the "tts" trigger or keyword. (2) Content needs to be spoken rather than read (multitasking, accessibility, podcast, driving, cooking). (3) Using pre-cloned…
natural-talk
A writing-style guide for producing natural-sounding Chinese conversations and creative fiction. It defines rules for avoiding artificial wording, unsupported claims, and unwanted changes to supplied text.
qwen-asr
Transcribe audio files using Qwen ASR. Use when the user sends voice messages and wants them converted to text.
lateral-thinking
Activate when: user says 'we've tried everything,' 'think outside the box,' 'we need a breakthrough,' 'challenge assumptions,' or 'how might we'; a team keeps cycling through the same 2-3 options; an industry assumption has gone unchallenged for years; conventional analysis keeps returning to the same answer. Do NOT…
cross-disciplinary-creativity
Activate when: 'we need a breakthrough idea', 'everything in our industry looks the same', 'we've exhausted all options in our field', 'how do we innovate in a way no competitor has tried', 'need to think completely outside the box'. Do NOT activate when: home domain has a known solution needing better execution; user…