FPV Immersive Video Prompting: skill pack for Hermes Agent

Hermes skill that plans camera routes for first-person AI video prompts, with image asset packs

What FPV Immersive Video Prompting does

The skill treats an FPV video as a movement to design rather than a picture to describe. It breaks each request into eight questions: who the camera is, where it starts, exactly how many main subjects there are, the order in which it passes them, whether each leg of the route is physically reachable, what happens at each stop, which identities and clothing must stay consistent, and what must never appear in the final frame.

It then picks one of two modes. Numbered stops suit indoor scenes with several characters, using small numbers on the first frame instead of a route line. Red-line path control suits aerial and world-scale routes, where a red line guides the camera but must vanish from the final video. For scenes with N main characters it generates a GPT Image asset pack: a numbered first frame, N character reference images and an optional clean first frame. It targets image-to-video tools such as Seedance, Kling, Runway and Veo.

Key features

Eight-question planning pass covering camera identity, start point, subject count and route order
Numbered-stop mode for indoor multi-character scenes
Red-line path control mode for aerial and map routes
GPT Image asset pack: numbered first frame, character references and a clean first frame
Prompts written for Seedance, Kling, Runway and Veo workflows

When to use it

Writing a 15-second video prompt where a cat walks past three people in a cafe
Planning a Seedance 2.0 fly-over that follows a route from a snowfield to a volcano
Keeping character faces and clothing consistent across a multi-subject shot

Who it is for: Creators who produce image-to-video AI clips and want controlled camera movement instead of loose scene descriptions.

How it fits with Hermes Agent

It is a Hermes skill (repository topic hermes-skill) installed into ~/.hermes/skills/creative/, and the README says the SKILL.md can also be used in Claude, Codex, Cursor and OpenCode.

How to install FPV Immersive Video Prompting

These commands are copied from the project's README. Check the repository for the latest steps before you run them.

git clone https://github.com/zhouwei713/fpv-immersive-video-prompting.git \
 ~/.hermes/skills/creative/fpv-immersive-video-prompting

Note: The main README is written in Chinese, with an English version linked as README_EN.md.

FAQ

What is FPV Immersive Video Prompting?

FPV Immersive Video Prompting is an agent skill that designs the camera route for first-person AI videos before writing the prompt. It outputs a video prompt and, where needed, a set of GPT Image prompts for the reference frames.

Does FPV Immersive Video Prompting work with Hermes Agent?

Yes, the README gives a Hermes install that clones the repository into ~/.hermes/skills/creative/. Restart Hermes or open a new session to use it.

How do I install FPV Immersive Video Prompting?

Clone the repository into ~/.hermes/skills/creative/fpv-immersive-video-prompting, then restart Hermes or start a new session. Without Hermes, you can place SKILL.md in another agent that supports skills or long prompts.