Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/sanky369/vibe-building-skillsnpx agentmods add skills/sanky369/vibe-building-skills/talking-headWrote 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/sanky369/vibe-building-skills/talking-head)<a href="https://agentmods.dev/skills/sanky369/vibe-building-skills/talking-head"><img src="https://agentmods.dev/badge/skills/sanky369/vibe-building-skills/talking-head/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/sanky369/vibe-building-skills/talking-head"><img src="https://agentmods.dev/badge/skills/sanky369/vibe-building-skills/talking-head.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00149 | $0.02139 |
| Opus 5.5 | $0.00060 | $0.00856 |
| Sonnet 5.5 | $0.00030 | $0.00428 |
| Haiku 4.5 | $0.00015 | $0.00214 |
Grade A, and why
talking-head 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 2mo 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 — 131 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Talking Head
Turn a message and an audience into a complete Talking Head Video Spec — one presenter archetype fully specified (styling, setting, lighting, framing), a timed hook/body/CTA script skeleton, and keyframe prompts ready for fal-ai/nano-banana-pro. The governing principle: credibility is cast, not scripted — the presenter's look, setting, and register must match what the audience already trusts for this kind of message (a polished studio expert selling "authentic user review" reads as an ad and dies). You produce the plan and still keyframes; voice, motion, and editing happen downstream.
When to use / when not to
- Use for any video fronted by a person: testimonials, UGC-style ads, educational/explainer videos, founder updates, product demos with a presenter, corporate training.
- Product-only motion (no person) →
skills/creative/product-video. - Programmatic/animated videos rendered in Remotion →
skills/creative/remotion-script-writer. - Static social assets →
skills/creative/social-graphics. - No brand style defined and consistency matters → run
skills/creative/creative-strategistfirst; its style guide feeds presenter styling.
Intake
Ask in one tight batch, only what's missing:
- Message and goal — what is the presenter saying, and what should the viewer do after (CTA)?
- Audience and platform — who watches, and where (TikTok/Reels, YouTube, landing page, internal training)?
- Trust mode — should this feel like a peer's recommendation (authentic/UGC) or an expert's word (authority/polished)?
- Constraints — real person to feature (founder, customer) or generated/hired presenter? Brand style guide from
skills/creative/creative-strategist?
Infer instead of asking: platform → length and framing; product category → default trust mode (consumer social ad → UGC; B2B/finance/health → authority). Don't stall: with message + audience known, state assumptions ("Assuming UGC style, 30s, vertical — say otherwise") and proceed.
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.
- 2mo ago First seen · 131 lines · 149 tokens per session scan A c86113e6e7d8
talking-head is a skill published in the GitHub repository sanky369/vibe-building-skills (31 stars, last pushed 3mo ago), licensed MIT. It adds 149 tokens to every session and 2,139 once invoked, about $0.0006 per session on Opus 5.5. 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 2026-08-30.
talking-head compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| talking-headsanky369/vibe-building-skills | Skill | 31 | 3mo ago | MIT | A | 149 | Claude Code, Codex |
| ads-youtubeAgriciDaniel/claude-ads | Skill | 11k | 5d ago | MIT | A | 84 | Claude Code |
| ads-launchAgriciDaniel/claude-ads | Skill | 11k | 5d ago | MIT | A | 45 | Claude Code |
| happiness-skillkangarooking/cangjie-skill | Skill | 12k | 1mo ago | MIT | A | 136 | Claude Code, Codex |
| setup-matt-pocock-skillsdevcxl/mattpocock-skills-zh | Skill | 530 | 3d ago | MIT | A | 43 | Claude Code, Codex |
| aiwg-regenerate-copilotjmagly/aiwg | Skill | 229 | yesterday | MIT | A | 25 | Claude Code, Codex |
| steward-prep-deliveryjmagly/aiwg | Skill | 229 | yesterday | MIT | A | 32 | Claude Code, Codex |