Scenario-labs/skills provides over 100 Agent Skills that teach AI coding agents (Claude Code, Cursor, Codex, Copilot, and 70+ others) to create production-ready assets via the Scenario MCP server (https://mcp.scenario.com/mcp). The collection covers images, video, audio, textures, skyboxes, 3D assets, and custom model training for games and creative verticals. Core skills include scenario for connecting to the MCP and running the generation loop, scenario-image for text-to-image with model choice, sizing, references, and masked inpainting, and scenario-multi-reference for composing one image from several references with leak checks. Video skills span scenario-video (text-to-video, image-to-video, lipsync, upscaling), scenario-seedance-music-video (beat-aligned shots, lyric transcription), and scenario-video-ads (brief to delivery with fidelity gates). Audio skills cover scenario-audio (music, SFX, voice, speech), scenario-minimax-music (vocal songs from tagged lyrics), and scenario-elevenlabs (TTS, music, SFX, dubbing). 3D skills include scenario-3d (text/image to mesh, multi-view reconstruction, GLB/FBX), scenario-video-to-motion (motion capture from video, auto-rigging, engine-ready exports), and scenario-walkable-room (interior photo to physics-enabled splat world). Expert tool families drive DCC software and game engines installed locally: ZBrush 2026 (9 skills, bridge proven live), Blender 5.2 (15 skills, 12 test suites pass headless on 5.2.1), Maya 2027 (11 skills), Unreal Engine 5.8 on macOS (10 skills via Editor Python and Epic's MCP server), Unity, Godot, and Roblox Studio. Installation uses npx skills add scenario-labs/skills with optional --skill flags or role-based commands from INSTALL.md; pinning via npx skills@1.7.4 add scenario-labs/skills#<commit>. Releases are managed by release-please with CHANGELOG.md and GitHub releases.