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 agentmods add skills/moses607/socialforge/youtube-title-labnpx skills add moses607/socialforge --skill youtube-title-labgit clone --depth 1 https://github.com/moses607/socialforgeWrote 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/moses607/socialforge/youtube-title-lab)<a href="https://agentmods.dev/skills/moses607/socialforge/youtube-title-lab"><img src="https://agentmods.dev/badge/skills/moses607/socialforge/youtube-title-lab.svg" alt="Measured on agentmods" height="20"></a>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 | $0.00126 | $0.01075 |
| Opus 5 | $0.00063 | $0.00537 |
| Sonnet 5 | $0.00025 | $0.00215 |
| Haiku 4.5 | $0.00013 | $0.00108 |
Grade A, and why
youtube-title-lab 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 — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.
YouTube Title Lab
On YouTube nothing else runs until the title/thumbnail combo wins the click — not the intro, not the retention edit, not the algorithm. The title and thumbnail are not two labels for the same idea; they are two halves of ONE curiosity gap. The thumbnail carries emotion and visual stakes (a face, a result, a contrast); the title carries the specifics and the promise. When they repeat each other you waste half your real estate; when they complete each other the viewer needs to click to close the loop. Every choice below optimizes the gap, then guarantees the video actually pays it off.
1. Generate 10-15 titles across 5 angles
Never brainstorm one angle. Force at least two titles per angle so you compare hooks, not phrasings:
- Curiosity — open a loop, withhold the payoff: "I Tried X for 30 Days. Nobody Warned Me."
- Result/proof — lead with the concrete outcome + number: "How I Got 10k Subs With 4 Videos."
- Contrarian — attack the consensus: "Stop Warming Up Before You Lift."
- How-to/utility — clean promise, front-loaded keyword: "Fix Slow WiFi in 60 Seconds."
- Story/stakes — personal, high-consequence: "This Mistake Cost Me $12,000." Constraints on every option: ≤60 characters (mobile truncates ~40-50), front-load the hook word in the first 1-3 words, one specific number or proper noun where possible, no title-case keyword stuffing.
2. Pair each finalist with a thumbnail-text concept (≤3 words)
Take your top 3-4 titles. For each, write the thumbnail text that completes — never echoes — it. Rule of division of labor:
- Title says the specific promise → thumbnail says the emotional stakes ("$12,000" title → thumbnail: "MY BIGGEST MISTAKE").
- Title asks the question → thumbnail shows the shocking answer-teaser, not the answer.
- Thumbnail carries ONE readable idea at arm's length on a phone: ≤3 words, high contrast, one focal face or object. If the title and thumbnail text say the same noun, kill one and reload. State, for each pair, the single curiosity gap it opens in the viewer's head.
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 · 58 lines · 126 tokens per session scan A 8ab47f625cd3
youtube-title-lab is a skill published in the GitHub repository moses607/socialforge (2 stars, last pushed 1mo ago), licensed MIT. It adds 126 tokens to every session and 1,075 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 2026-08-31.
Other skills, from other repositories
blog-notebooklm
Query Google NotebookLM notebooks for source-grounded, citation-backed answers from user-uploaded documents. Manages notebook library, handles Google authentication, and supports smart discovery. Works standalone via /blog notebooklm or internally from blog-write and blog-researcher for source-grounded research…
paper-explainer
科研论文解读：解析 arXiv/PDF 的公式与图表，提炼问题、贡献、方法、关键图和结论，再产出 B站/视频号解读视频或知乎/公众号图文。 当用户说“论文解读、讲论文、论文转视频/图文、科研科普、arXiv、学术视频”时使用。 本 SKILL 从论文做内容；video-to-article 从视频做图文，doc-convert 只转换文档格式。.
infographic
将数据或文字内容转化为可视化信息图，支持静态（AntV）和动画 GIF 两种模式。当用户需要制作信息图、数据可视化、流程图、对比图、动画图表、GIF 图表、思维导图、SWOT 分析图时调用。本地渲染信息图/GIF 动画；要单张静态图片 URL 用 chart-visualization，要 CSV/JSON→整页报告用 data-report.
data-report
把 CSV、Excel 或 JSON 数据生成包含 KPI、图表和洞察的完整可视化报告页。当用户说“数据报告、CSV/Excel 转报告、做 KPI 看板、生成可视化报告页”时使用。单张图表用 chart-visualization；信息图或 GIF 动画图表用 infographic。.
doc-convert
把 Markdown 文稿排版并转换为 HTML、可打印 PDF 或长图 PNG。当用户说“Markdown/MD 转 HTML/PDF/图片、文章导出长图、MD 排版/渲染”时使用。仅处理 Markdown；DOCX/PPT 不在范围内，层级大纲转脑图用 mindmap。.
descript
The Descript craft skill — edit talk content (podcasts, interviews, talking-head video) by editing the transcript instead of the timeline. Use when someone wants to edit in Descript, edit a podcast or interview, remove filler words/silences, clean up audio (Studio Sound), fix a flubbed word without re-recording…