kanban-video-orchestrator

Plan and monitor multi-agent video production pipelines coordinated through an Anakot kanban board.

Updated Jun 7, 2026
One-click install
npx skills add https://github.com/Chensihakniroth/ANAKOT-AGENT --skill kanban-video-orchestrator-chensihakniroth
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: kanban-video-orchestrator
Source: https://github.com/Chensihakniroth/ANAKOT-AGENT/tree/main/optional-skills/creative/kanban-video-orchestrator
Command: npx skills add https://github.com/Chensihakniroth/ANAKOT-AGENT --skill kanban-video-orchestrator-chensihakniroth

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pyyaml, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve? Producing a video that mixes styles, tools, and roles (writing, rendering, voiceover, editing) is hard to coordinate manually. This Skill decomposes any video request into a team of specialized Anakot agent profiles whose work is routed, tracked, and reviewed through a kanban board. ## Core Features & Use Cases - Adaptive Discovery & Briefing: Asks targeted per-style intake questions (narrative, marketing, music video, explainer, ASCII, installation) and produces a structured brief.md that acts as the production contract. - Team Design & Setup Generation: Composes 4-7 role profiles (director, writer, renderers, voice-talent, editor, reviewer) and generates a setup.sh that creates profiles, SOUL.md files, workspace, and the initial kanban task via scripts/bootstrap_pipeline.py. - Execution Monitoring: Polls the kanban with scripts/monitor.py to detect stuck tasks, missing heartbeats, overtime renders, and repeated retries, with documented intervention recipes. - Use Case: Ask for a 30-second product teaser; the Skill scopes the brief, designs a team with a motion-graphics renderer and ASCII renderer, fires the kanban, and helps you monitor until final.mp4 passes review. ## Quick Start Ask the agent to plan and run a kanban-based production pipeline for the video you want to make, describing its length, style, and target platform.

Frequently Asked Questions about kanban-video-orchestrator

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I set up a multi-agent video production pipeline?▼

Define a plan.json with your team roles, scenes, audio, and deliverables, then run scripts/bootstrap_pipeline.py to generate setup.sh. The script creates Anakot profiles, writes SOUL.md files, builds the workspace, and fires the initial kanban task to the director.

What rendering tools can a video pipeline use per scene?▼

Scenes route to whichever skill fits the beat: ascii-video, manim-video, p5js, comfyui, touchdesigner-mcp, blender-mcp, pixel-art, or baoyu-comic. External APIs like Runway, Kling, and ElevenLabs handle image-to-video and voiceover via the terminal toolset.

When should I not use a kanban video pipeline?▼

Skip it for one-shot conversions like mp4-to-GIF (use ffmpeg directly), static images or audio-only artifacts (use the matching specific skill), or single-tool projects that one existing skill covers cleanly, such as a pure ASCII video.

How do I monitor a running video production kanban?▼

Run scripts/monitor.py with your project tenant to poll the board and surface stuck tasks, missing heartbeats, overtime renders, and repeated retries. You can also use anakot kanban watch for live events or anakot dashboard for a visual board.

Why do kanban video tasks fail with missing credential errors?▼

External APIs like ElevenLabs, Runway, and OpenRouter require keys in ~/.anakot/.env or macOS Keychain. The setup script's check_key helper verifies required keys before firing the kanban and aborts cleanly if any are missing.

Can multiple renderer profiles share project files?▼

Yes, all profiles share one workspace when every kanban_create call passes workspace_kind="dir" and the absolute project workspace_path. Using worktree isolation instead breaks artifact sharing between profiles.