montaj

You MUST use this whenever the user asks for video editing work. Use it when video-related tasks are brought up. Editing, analyzing video, or transcribing videos

Find and download real images from the web — people, logos, brand/event stills, B-roll — to add as overlay image cards or project assets. Load when the prompt asks to source or insert images (e.g. 'add a photo of X', 'find images of the IPO', 'pull a shot of the factory').

Agent-authored workflow task: analyze transcripts across all clips, pick ONE best take per script section, discard all others. Load this when you hit montaj/select_takes in a workflow.

Write a custom JSX overlay component and add it to the project's overlay track.

Agent-authored workflow task: transcribe-and-sample one long-form horizontal source, pick N self-contained clip windows, and fan each out into its own vertical (9:16) clip project with the chosen framing. Load this when you hit montaj/find_clips in a workflow.

The shared vocabulary that domain skills use to name Montaj operations. Domain skills phrase every Montaj interaction as one of these verbs; an interface skill defines how each verb is actually performed. Load this to learn the canonical verbs before authoring or reading a domain skill.

The native interface — how the _contract verbs are performed when Montaj runs locally, either via the HTTP server (montaj serve) or the headless CLI. The dispatcher loads this after it detects HTTP or CLI mode; domain skills never read it directly. Covers step invocation, project read/save, file read/write, logging, GET-before-save discipline, workspace resolution, and image-step credentials.

Agent-authored workflow task: decide what overlays to write, author the JSX, and add them to the project's overlay track. Load this when you hit montaj/overlay in a workflow.

Agent-authored workflow task: build slides for a carousel project — pick aspect-aware images via generate_image, position image and overlay elements on each slide, set base colors, and render to PNGs. Load this when you hit montaj/carousel in a workflow.

Author carousel text overlays whose font/size/color/etc. the operator can edit from the editor toolbar. Load when authoring text overlays inside a carousel project.

Orientation skill for new agents and users. Explains what Montaj is and how it works. Use it when the user asks questions about montaj or is getting started

Generation orchestration for approved ai_video storyboards: dispatch mode selection, scene generation via kling_generate, audio assembly, regenQueue processing. Load after storyboard approval.

Agent-authored director skill for the ai_video workflow: story clarification, storyboard writes, approval gate. Load this when you hit montaj/ai-video-plan in a workflow or encounter a project with projectType: 'ai_video'.

Quality evaluation rubric + retry loop for ai_video generated clips. Load after generating scenes in Phase 6 to evaluate and optionally regenerate clips that fail quality checks.

Shot scale, camera move taxonomy, and cinematography rules for ai_video scene planning. Load when writing storyboard scenes in Phase 1.

Agent-authored workflow task: choose a render path (ffmpeg vs JSX), build project.json overlay entries from lyrics_sync output, and set up audio. Load this when you hit montaj/lyrics-video in a workflow.

Parallel execution patterns for montaj — load when workflow has multiple clips or foreach steps

Agent-authored task: build animation-only sections from scratch using opaque overlays. Load when the agent hits montaj/animation-sections in a workflow.

Loaded after a draft is ready for back-and-forth editing: cutting, overlay adjustments, timing fixes. Not a pipeline step — this is the interactive editing reference.

Create or edit Montaj workflows. Reasons about step dependencies to produce a workflow JSON with correct `needs` fields for parallel execution.

Agent-authored workflow task: generate waveform images per clip and reason visually about silence vs. speech to produce trim specs. Use when waveform_trim's fixed threshold fails due to inconsistent background noise.

MCP server mode for montaj — load when running as an MCP client (e.g. Claude Desktop)

Create or update a creator style profile — analyzes video content to extract editing style, pacing, color palette, and aesthetic direction for use in all future editing decisions.