Mandatory for every Nomi corrective change: user-reported bugs, regressions, CI-only failures, flaky tests, performance or security defects, review/audit findings, and compatibility failures in any production path. Classify oneoff versus recurring before implementation. Recurring and high-risk repairs require a…
A set of project instructions for aqm857886159/Nomi, covering its structure, commands, engineering rules, and checks before work, completion, or pushing changes.
A project instruction file for using Claude Code with the Nomi project, including project rules, commands, and checks before changes are finished or pushed.
A step-by-step plan for turning product copy and selling points into a short brand video. The structure is a three-second hook, key benefits, use cases, and a call to action, with review points between stages.
A guide to staging non-combat action scenes such as parkour, chases, climbing, stunts, and falls. It covers movement vocabulary, interaction with the environment, pacing, FPV chase shots, and physical constraints.
A guide to visual storytelling for video, covering shot sizes, composition, camera movement, lighting, focus range, color temperature, and lens characteristics. It also explains how to turn these choices into video prompts.
A method for keeping characters, props, settings, and visual style consistent across multiple AI-generated video shots. It uses reference images, scene state notes, and links between shots.
A reference guide to historical Chinese costume drama, including clothing, hairstyles, makeup, buildings, etiquette, ceremonies, and period details. It also covers useful points for image and video prompts.
A checklist for reviewing reference images that define a character or scene. It gives step-by-step checks, a Go/No-Go decision, and feedback wording for requesting changes.
A guide to directing non-fight performances through facial expressions, body language, eye movement, silence, breathing, pauses, and repeated character habits. It turns emotions into visible physical actions for AI video prompts.
A handbook for turning a planned camera shot into clear video-generation wording. It covers camera movement, focus, combined shots, unwanted wording, and translating abstract actions or emotions into physical actions.
A guide to planning sound for stories, including background noise, effects, music, silence, and recurring musical or sound themes. It is for planning and editing reference; generated video prompts do not contain audio.
A guide to staging scenes with three or more people, including where they stand, look, and draw attention. It also covers keeping multiple characters consistent in AI-generated video.
A directing guide that combines the styles of Kishi Otomo and Edgar Wright for stylish action and heist scenes. It matches camera movement, editing pace, and sound-image timing to scene types such as chases, fights, break-ins, emotional dialogue, and fashion entrances.
A guide to joining separately generated video shots with creative transitions such as occlusion, whip pans, fades, match cuts, and objects passing the camera. It also treats transitions as tools for changing scenes, time, characters, or mood.
A Chinese-language playbook for turning a story outline or novel excerpt into a short dramatic video with characters, plot turns, and planned shots. It works in stages, pausing for review between script, character, scene, image, video, and timeline work.
A release playbook for preparing the media materials for a Nomi desktop app update. It covers checking version facts and public examples, planning and generating the story, and handing over finished videos, update materials, social copy, and local quality checks in two languages.
A translator that turns a skill, workflow document, or short requirement into a usable Nomi skill. It maps available tools, declares capabilities, and records gaps that cannot be supported.
An AI writing assistant for ordinary writing, scripts, prompts, and storyboard descriptions. It suggests edits through a document action process, and changes are written only after the user confirms them.
A Chinese-language planning skill that extracts main characters and important locations from a script and creates a card for each. It adds identity-board and scene-board prompts to a generation canvas to keep characters and settings consistent.