Video Generation

Content updated.

Content updated.

Content updated.

Display title updated.

- Renamed the skill from "video-cog" to "video-generation-cellcog". - Updated documentation and references to reflect the new skill name. - Removed the "skill-card.md" file. - No functional changes to core capabilities or usage; all major features and usage examples remain the same.

- Added explicit requirements for system binaries and environment variables (python3, CELLCOG_API_KEY) in the metadata section. - No other content changes; documentation and usage instructions remain the same.

- Updated SKILL.md description for improved clarity and expanded video use cases, including social media reels and YouTube content. - Adjusted usage instructions, clarifying agent compatibility for OpenClaw and other agents. - Streamlined sample prompts and capability descriptions to better outline supported video types. - No code or API changes; documentation only.

- Updated the skill description for improved clarity and detail in the SKILL.md file. - Added a more explicit example for initializing CellCogClient in usage instructions. - Enhanced feature highlights to more clearly list supported video types. - No functional or code changes; documentation updates only.

**Changelog for video-cog v1.0.10** - SKILL.md significantly condensed and streamlined for clarity and brevity. - Simplified feature and capability descriptions in the overview and usage instructions. - Updated usage examples and code snippets for improved guidance. - Pruned introductory and motivational text, focusing on actionable information. - Added new "Tips for Better Videos" section to assist with prompt crafting. - Retained key specifications, video types, and generation caveats for transparency.

- Expanded SKILL.md with detailed descriptions of supported video types, genres, and example prompts. - Added clear instructions and best practices for lipsync/spokesperson videos and agent team usage. - Provided specific prompt templates for marketing, explainer, educational, and spokesperson videos. - Clarified expectations: video generation remains experimental, with results varying by prompt and complexity. - Updated and broadened video specification details (aspect ratios, styles, audio options). - Emphasized the multi-model, multi-agent orchestration that underpins the skill’s functionality.

**Summary:** Major documentation simplification and focus in SKILL.md. - SKILL.md rewritten for clarity and brevity; now highlights core features, capabilities, and use cases in a more concise format. - Added explicit technical details: supported video durations, aspect ratios, audio modes, and output format (MP4). - Documented internal pipeline: animation, image generation, audio, video editing. - Expanded related skills section for broader use cases. - Removed lengthy examples and deep-dives on video styles, focusing instead on core functionality and quick understanding.

- Updated connection instructions in the SDK setup section to clarify usage for OpenClaw agents (now uses notify_session_key for fire-and-forget tasks) and for all other agents. - Streamlined references to the cellcog documentation. - No changes to code or functionality; documentation only.

video-cog 1.0.6 - Updated usage instructions to use a concise "Quick start" code sample instead of the previous fire-and-forget example. - Added a reference directing users to the cellcog skill for full SDK API documentation and usage patterns. - Clarified prompt parameters and usage by simplifying initial setup steps. - No changes to functionality—documentation update only.

- Updated the skill description for clarity and conciseness. - Added `homepage` and expanded `metadata` fields (including supported operating systems). - No core functionality changes; documentation improvements only. - The skill file now provides a clearer overview and more accessible information for users.

- Added a new section setting clear expectations about the unpredictability, risks, and learning curve of long-form AI video generation with CellCog. - Updated video specifications: maximum video duration now clearly stated as 4 minutes (was "5+ minutes"). - All other instructions and usage guides remain unchanged.

- Added `author` and `dependencies` fields to metadata for clarity and package management. - Updated prerequisite instructions to refer to the `cellcog` skill by package name. - No functional or interface changes; documentation improvements only.

- Updated description and introduction to emphasize long-form AI video production (up to 4-minute videos) and multi-agent model orchestration with CellCog. - Added highlights of production steps: scripting, filming, voice, lipsync, scoring, editing—all automated. - Clarified capability scope: marketing, demos, explainers, educational, spokesperson, UGC, news, and more. - Stressed CellCog’s orchestration of 6-7 foundation models in single-prompt workflows. - No changes to usage patterns, prompts, or API; this update is primarily documentation and positioning.

- Updated the quick start guide to use the new `create_chat` pattern with `notify_session_key` and `task_label` for fire-and-forget video generation. - Added an OpenClaw metadata emoji to improve discoverability. - Clarified the prerequisite section for SDK setup. - Minor edits to streamline language and remove redundant information about session patterns. - No changes to supported video types, examples, or general usage instructions.

video-cog 1.0.0 – Initial Release - Launch of CellCog-powered AI video generation skill. - Supports creation of marketing, explainer, educational, documentary, cinematic, UGC, and news-style videos. - Enables lipsync and spokesperson videos with AI-generated voice and character presentations. - Includes comprehensive usage guide, example prompts, and video specifications. - Requires CellCog mothership skill for setup and API integration.