dsh-fight-scene-director
Run the following command in DeepSeek Harness:
dsh plugin install hatsuyuki0103/dsh-fight-scene-director
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install hatsuyuki0103/dsh-fight-scene-director to install the plugin from https://github.com/hatsuyuki0103/dsh-fight-scene-director, then restart DeepSeek Harness to activate the fight-scene-director skill.
About this plugin
Producing AI-video fight scenes is hard not because you cannot describe characters and weapons, but because action must have causal logic and the camera must track spatial motion. dsh-fight-scene-director takes your character sheets, weapons, abilities, scene notes, and target duration, then emits a structured Chinese fight prompt aimed at Seedance 2.0, Seedance 2.5, and MiniMax H3, with action causality and spatial movement always ahead of shot selection.
It goes beyond a single text prompt. The plugin generates three image-level directives: an initial-position map, a character-trajectory map, and a character-plus-camera dual-trajectory map, turning staging, motion paths, and camera follow into frame-by-frame instructions. Encounter types span melee, ranged, spell, group, summon, and transformation fights; durations of 15 s, 30 s, or custom are supported, followed by an adjustable rhythm panel. Critically, continuity is enforced across cuts: character positions, weapon states, ability activations, and environmental destruction stay coherent so that a strike landing on the left in one shot does not jump to the right in the next.
This plugin is built for creators already using Seedance or MiniMax H3 to produce action short films, game trailers, or anime-style fight sequences who keep hitting the wall of scattered action and broken spatial logic. Inside DeepSeek Harness, a single plain-language request such as design a 30-second sword-duel fight scene is enough to trigger the full pipeline, with no need to hand-write a structured prompt.
Use Cases
- Generate causality-driven fight prompts with spatial trajectories for Seedance or MiniMax H3
- Keep character blocking, weapon state, and environmental destruction coherent across cuts
- Convert character sheets, stage notes, and target duration into position maps and dual-trajectory directives
Best For
- Creators producing action shorts or game trailers with Seedance or MiniMax H3
- Independent animators and filmmakers who cannot pin down scattered action or broken staging
- Pre-production planners and storyboard artists seeking structured shot language and rhythm panels
Related Plugins
A unified suite combining hot runtime injection, task-aware thinking-mode routing, and a graded session protocol with red-team gates to sustain model diligence across long-horizon inference.
ModLens is a vision plugin for DeepSeek Harness that gives text-only models sight by reading images pasted directly into chat, with zero-config setup and multiple vision engines.
On-demand vision for text-only DeepSeek Harness agents: built-in free keyless vision chain and 14 vision tools, routing image turns as tool calls to vision models with pixel fidelity, no Python needed, one-command install.
Give text-only models in DeepSeek Harness eyes, enabling image Q&A, long-screenshot OCR, UI restoration, and GUI visual tasks.