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 leynos/visual-storytelling-skills --skill kling-3-0-deep-divegit clone --depth 1 https://github.com/leynos/visual-storytelling-skillsWrote 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/leynos/visual-storytelling-skills/kling-3-0-deep-dive)<a href="https://agentmods.dev/skills/leynos/visual-storytelling-skills/kling-3-0-deep-dive"><img src="https://agentmods.dev/badge/skills/leynos/visual-storytelling-skills/kling-3-0-deep-dive/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/leynos/visual-storytelling-skills/kling-3-0-deep-dive"><img src="https://agentmods.dev/badge/skills/leynos/visual-storytelling-skills/kling-3-0-deep-dive.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 | $0.00142 | $0.03336 |
| Opus 5 | $0.00071 | $0.01668 |
| Sonnet 5 | $0.00028 | $0.00667 |
| Haiku 4.5 | $0.00014 | $0.00334 |
Grade A, and why
kling-3-0-deep-dive 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 12d 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 — 336 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Kling 3.0 Deep Dive
Use this skill when shot-specifier routes a shot to kling3_0 or when
video-generator is about to submit Kling 3.0 jobs through the Higgsfield Model
Context Protocol (MCP).
Kling 3.0 is best treated as a scene-directed cinematic video model. It responds well to shot labels, camera behaviour, physical motion, clear timing, and simple subject/environment constraints. Think like a director and camera operator: what is the shot, what moves, how does the camera react, what does the audience hear, and where does the shot end?
Live-Schema Rule
Before any production generation, inspect the live Higgsfield MCP schema through
video-generator. Public guides disagree on exact model IDs, quality modes, text
capabilities, Elements support, Motion Control support, native audio options, duration
ranges, and media roles.
Use the limits below as planning defaults only. If the live MCP schema is narrower, follow the live schema. If the selected Kling route cannot accept the required anchors, Elements, motion references, or audio parameters, stop and ask for a production decision.
S01 session 2 observed a narrower Higgsfield MCP surface than public Kling guides
describe: Kling accepted start_image and end_image, did not expose a generic
reference-image role, did not expose sound, did not expose cfg_scale, and downloaded
16:9 outputs at 1344x768. Treat Elements, Motion Control, CFG, sound toggles, and
labelled resolution choices as schema-gated features, not assumptions.
When Kling 3.0 Is The Right Route
Prefer Kling 3.0 when the shot needs:
- explicit scene structure or 2-6 shot coverage in a single generation;
- clean cinematic camera movement: pans, tracks, dollies, orbits, reveals, drone moves;
- physically grounded movement, environmental interaction, impact, or macro texture;
- landscape establishing shots, machine/drone POV, exteriors, product hero moves, or camera-motion-led b-roll;
- start/end frame control or transition shaping;
- Elements or Motion Control when the live tool exposes them;
- native ambient audio, sound effects, or structured dialogue when the scene can be labelled unambiguously.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 12d ago First seen · 336 lines · 142 tokens per session scan A 63c137ef2c96
kling-3-0-deep-dive is a skill published in the GitHub repository leynos/visual-storytelling-skills (6 stars, last pushed 1mo ago), licensed ISC. It adds 142 tokens to every session and 3,336 once invoked, about $0.0007 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
wavespeed
Generate or edit AI media (image, video, audio, 3D) by calling the wavespeed CLI on the user's machine. Use whenever the user asks to create, edit, animate, upscale, or transform a visual asset, generate audio/TTS/music, or produce marketing creatives. Every model on the WaveSpeed platform is one wavespeed run call.
ai-hive-simulate-ali-abdaal
A research and content-planning workflow that studies Ali Abdaal’s public work on AI, technology startups, products, business, and teaching. It turns those observations into original material and advisory analysis without impersonating him.
ai-hive-simulate-aravind-srinivas
A research and content-planning workflow that studies Aravind Srinivas’s public work on AI, technology startups, products, business, and teaching. It turns those observations into original material and advisory analysis without impersonating him.
ai-hive-simulate-baoyu-xp
A research and writing workflow that studies Baoyu XP's public methods for explaining AI, technology, products, digital tools, and engineering. It turns those general methods into original content and does not impersonate Baoyu XP.
ai-hive-simulate-berryxia-ai
A research and writing workflow that studies Berryxia AI's public methods for AI tools, agents, AI programming, visual generation, products, and business use. It adapts those methods into original work without imitating the person's identity.
ai-hive-simulate-dario-amodei
A research and content-planning workflow that studies Dario Amodei’s public material about AI, technology companies, products, and management.