OpenParable
Skip to content
← Back to skills

Reference Driven Video Prompting

Documents
reference-driven-video-prompting

Reference Driven Video Prompting is an Agent Skill for documents published in the nolanx-ai/nolanx.ai repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Use for image-to-video, reference-video continuity, keyframe-driven prompting, camera control, and consistency workflows based on official Runway, Kling, and character/style reference patterns.

View on GitHub ↗
QUICK START

How to use this skill

Bring this guide into your coding agent with a prompt tailored to the tool you use.

  1. Open your project in Codex.
  2. Copy and paste the prompt below.
  3. Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/nolanx-ai/nolanx.ai/blob/HEAD/skills/reference-driven-video-prompting/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/reference-driven-video-prompting, but wait for my approval before writing files or running scripts.

Copying this prompt does not install or run the skill. Review third-party files before use.

Reference Driven Video Prompting

Use this skill when the workflow depends on image inputs, prior videos, keyframes, or reference-based consistency.

Source Basis

  • Runway Gen-3 prompting, keyframes, and camera control docs.
  • Kling image-to-video prompt pattern.
  • Midjourney character/style reference concepts for identity/style persistence.

Reference Hierarchy

  1. identity reference
  2. world/location reference
  3. previous formal shot continuity
  4. style reference
  5. motion prompt

Prompting Rules

  • With an input image, focus text on desired motion and camera behavior rather than re-describing the full image.
  • Use direct, descriptive motion language; avoid conceptual chatter.
  • Separate subject motion from camera motion.
  • When using keyframes, define what must transform between first/middle/last frame.

Consistency Rules

  • Character references preserve face, hair, clothing silhouette, and recurring identity traits.
  • Style references preserve palette, texture, light behavior, and rendering attitude.
  • Previous video references preserve action phase, movement direction, and world continuity.

Camera Control Rules

  • Choose one dominant camera move per shot unless complexity is essential.
  • If the subject moves toward camera, note whether the camera compensates, holds, or retreats.
  • Do not overload a shot with contradictory camera verbs.

Good Motion Language

  • sways gently in wind
  • advances in heavy armor with restricted weight
  • camera slowly pushes in as expression hardens
  • orbit tightens during transformation

Avoid

  • vague commands with no motion target
  • multiple unrelated action verbs in one short shot
  • re-randomizing identity when continuity references already exist