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 linkprint/local-ai-movie-workspace --skill h3-micro-fpv-impossible-one-takegit clone --depth 1 https://github.com/linkprint/local-ai-movie-workspaceWrote 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/linkprint/local-ai-movie-workspace/h3-micro-fpv-impossible-one-take)<a href="https://agentmods.dev/skills/linkprint/local-ai-movie-workspace/h3-micro-fpv-impossible-one-take"><img src="https://agentmods.dev/badge/skills/linkprint/local-ai-movie-workspace/h3-micro-fpv-impossible-one-take/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/linkprint/local-ai-movie-workspace/h3-micro-fpv-impossible-one-take"><img src="https://agentmods.dev/badge/skills/linkprint/local-ai-movie-workspace/h3-micro-fpv-impossible-one-take.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 | $0.00088 | $0.00911 |
| Opus 5 | $0.00044 | $0.00456 |
| Sonnet 5 | $0.00018 | $0.00182 |
| Haiku 4.5 | $0.00009 | $0.00091 |
Grade A, and why
h3-micro-fpv-impossible-one-take 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 — 47 lines — stays where its author put it; the contents beside it link to each section on GitHub.
H3 Micro FPV Impossible One-Take
Direct one continuous micro-camera path with legible geography, tactile scale, and accelerating but collision-free motion.
Route the Request
- For prompt writing or refinement, return text only.
- Read references/style-blueprint.md and use $h3-prompt-writing for exact H3 field order, timing, audio, and negatives.
- Invoke $h3-video-generation only when the user explicitly asks to generate or render.
- Use T2VA by default. Use I2VA only when one supplied opening frame actually matches the opening camera position. When a source portrait supplies identity, pose, prop, or room authority but the flight begins elsewhere, use native Ref2VA and bind the ordered files with
<Picture N>or<Video N>. - For the selected neo-noir portrait route, read the exact spatial and focus choreography in the reference file. Use 15 seconds for the full route or its explicit 10-second safe demo compression.
- When the user supplies three images of the same subject in different poses, use the anchor-flow route in the reference: the poses are spatial waypoints, not frozen destinations or cuts.
Build the Flight
- Define camera size, lens height, speed, and the first surface at macro scale.
- Map a single connected route through three to five spatial gates or around one stable subject. Each next prop, limb, opening, or landmark must be visible or foreshadowed before the camera reaches it.
- Anchor scale repeatedly with fibers, dust, droplets, scratches, fasteners, seams, or tool marks.
- Choreograph focus as part of the route. Every rack-focus handoff must be motivated by camera distance and connect named foreground, midground, and background anchors.
- Use natural foreground occlusions from props, fabric, hair, smoke, heels, tools, or architecture, but remain spatially coherent and never conceal a cut.
- Use speed ramps around obstacles, then one short breath before the final reveal. Keep forward momentum continuous.
- Specify exact clearances and parallax. Do not teleport, clip through solids, change lens size abruptly, cross a weapon muzzle, or hide cuts in motion blur.
- Add airflow reactions and synchronized pass-by sounds for every near miss.
- End on one strong scale reversal, destination reveal, or stable hero composition, not a random crash.
What ships with it
2 files 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 · 47 lines · 88 tokens per session scan A 652e70cf2d51
h3-micro-fpv-impossible-one-take is a skill published in the GitHub repository linkprint/local-ai-movie-workspace (1 stars, last pushed 14d ago), licensed MIT. It adds 88 tokens to every session and 911 once invoked, about $0.0004 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
tape-opens
Stack-unwind review of this conversation - a tree of main tasks done (struck through) and every open left behind at the depth it was abandoned, sized by the thread's measured shape, never by feel. Optional argument - a .jsonl path or session id to review a different session.
higgsfield-troubleshoot
Use when a Higgsfield generation fails, produces poor quality, looks wrong, doesn't match the prompt, or the user needs to fix or improve an output.
higgsfield-mixed-media
Use when the user asks about Mixed Media, wants to apply artistic preset styles to an image (Noir, Sketch, Paper, Canvas, Particles, Neon, etc.), combine multiple artistic treatments, or create stylized non-photorealistic outputs.
higgsfield-motion-design
End-to-end motion-design / animated-ad creation flow on Higgsfield via the MCP connector. Use when the user wants to create motion design, animate a logo, make a video from an image, build an animated ad or brand promo, turn a product into motion, or says 'make a motion', 'motion design', 'animate this', 'make a video…
higgsfield-vibe-motion
Use when the user asks about Vibe Motion, motion graphics, kinetic typography, animated text, brand animations, logo animations, infographic animation, presentation animation, data-driven video, or prompt-to-motion-graphic workflows. Vibe Motion is fundamentally different from video generation — it creates…
company-brain
How an agent should operate as one mind inside a shared Company Brain built on Caura — recall before acting, obey fleet keystones, reuse and publish skills, and report outcomes so every task compounds across the whole organization. Use this whenever you work as part of a Caura-connected team or fleet and your work…