← Back to Skills Marketplace
13992
Downloads
10
Stars
101
Active Installs
13
Versions
Install in OpenClaw
/install image-cog
Description
AI image generation and photo editing powered by CellCog. Text-to-image, image-to-image, consistent characters, product photography, reference-based generati...
Usage Guidance
Install only if you trust CellCog and the cellcog Python package. Use a revocable API key, keep CELLCOG_API_KEY out of prompts and logs, and do not submit private photos, confidential brand assets, or regulated data unless CellCog's provider routing and data-handling terms are acceptable.
Capability Analysis
Type: OpenClaw Skill
Name: image-cog
Version: 1.0.12
The image-cog skill bundle provides documentation and usage examples for an AI image generation service powered by the cellcog Python library. It outlines capabilities for text-to-image generation, editing, and character consistency across multiple models. The bundle contains no executable code, malicious instructions, or evidence of data exfiltration, and its behavior aligns with its stated purpose.
Capability Assessment
Purpose & Capability
The skill consistently describes CellCog-powered AI image generation and editing, including text prompts, reference images, and routing to Google, OpenAI, and Recraft models; these capabilities match the stated purpose.
Instruction Scope
The runtime instructions are usage examples for CellCog chat/image generation and setup. I found no hidden commands, prompt-injection text, destructive actions, local data harvesting, or unrelated account actions.
Install Mechanism
The bundle contains only SKILL.md and no executable scripts. Setup asks users to install the CellCog skill or run pip install -U cellcog, which is purpose-aligned but depends on trusting the external Python package.
Credentials
Requiring python3 and CELLCOG_API_KEY is proportionate for an API-backed image service. Prompts and uploaded/reference images are expected to leave the local environment for CellCog and routed providers, so users should avoid sensitive content unless approved.
Persistence & Privilege
The artifact does not request persistence, background startup, elevated privileges, protected-path writes, credential scraping, or broad filesystem indexing.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install image-cog - After installation, invoke the skill by name or use
/image-cog - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.12
- Added environment and binary requirements to metadata: `python3` and `CELLCOG_API_KEY` are now specified as dependencies.
- No functional or user-facing changes; documentation metadata improved for deployment clarity.
v1.0.11
- Expanded description to highlight support for brand assets, stickers, comics, and GIFs.
- Clarified agent usage instructions: now specifies "All agents except OpenClaw" for one code sample.
- No changes to code, features, or major usage – documentation and wording improvements only.
v1.0.10
- Revised and expanded the skill's description for greater clarity and detail.
- Updated SDK usage examples, including new agent provider options.
- Improved and clarified image generation features, use cases, and example prompts.
- Enhanced documentation for image specifications, styles, and recommended use cases.
- Added instructions for both fire-and-forget and blocking agent modes.
v1.0.9
- Shortened and clarified the skill description for improved readability.
- Streamlined usage instructions, separating OpenClaw and other agents.
- Removed redundant and verbose onboarding and prerequisite information.
- Kept all key examples, model details, image types, and prompt tips intact.
- No changes to core functionality or capabilities; documentation update only.
v1.0.8
**image-cog v1.0.8 changelog**
- Completely overhauled documentation in SKILL.md for clarity, use cases, and guidance.
- New sections detailing available models (Gemini 3.1 Flash, GPT Image 1.5, Recraft) and when each is used.
- Expanded examples for single images, product photography, sets, character consistency, and reference-based generation.
- Added practical table summaries for aspect ratios, formats, and when to use agent team mode.
- Improved prompt-writing tips and concrete prompt samples to help users get better results.
v1.0.7
**Summary: Major documentation and feature list update for broader capabilities.**
- Expanded description and examples for image editing, vector, icon, and transparent background support.
- Output formats clarified: PNG, JPG, WebP, SVG, with per-model best-use notes.
- New section on internal model routing and reference image handling.
- Added chat mode recommendations for single versus batch/multi-image scenarios.
- Linked to related skills for branding, memes, 3D, GIF, and sticker generation.
- Broader and clearer language for use cases and feature highlights.
v1.0.6
- Added guidance on using OpenClaw "fire-and-forget" agent sessions for long image tasks
- Clarified usage of `notify_session_key` and agent differences in code examples
- Improved SDK usage instructions and removed references to deprecated delivery modes
- No changes to skill functionality or API, documentation only
v1.0.5
- Updated documentation to clarify quick start and SDK usage.
- Shortened and simplified the code example for starting image generation.
- Added direction to consult the cellcog skill for full API/details.
- No changes to skill features or logic; documentation only.
v1.0.4
- Expanded description to highlight image-to-image, photo editing, and social media content generation.
- Added supported operating systems and a homepage link to metadata.
- Improved clarity in usage instructions and feature lists.
- Strengthened emphasis on multi-model capabilities and routing.
- Updated skill documentation for broader and clearer coverage of features.
v1.0.3
**Now supports multi-model image generation, with automatic provider routing for different image tasks.**
- Added a new section detailing which image generation models are used and their roles: Nano Banana 2 (default), GPT Image 1.5 (for transparent images), and Recraft (vector/icon generation).
- Clarified that CellCog agents intelligently select the best model for your request, and users can request a specific model in the prompt.
- No API pattern or usage changes.
- All other documentation remains unchanged except for new model explanation.
v1.0.2
- Added explicit author (CellCog) and dependencies ([cellcog]) metadata to skill manifest.
- Clarified prerequisite to require the `cellcog` skill (previously “CellCog mothership skill”).
- Minor improvements to formatting and dependency instructions for clarity.
- No changes to feature set or functionality.
v1.0.1
- Added an emoji to the skill metadata for improved identification.
- Updated example usage for image generation to show the new fire-and-forget pattern using `create_chat` (v1.0+).
- Clarified SDK setup instructions and removed references to the deprecated `sessions_spawn` pattern.
- Small documentation improvements for clarity and accuracy.
v1.0.0
- Initial release of image-cog skill: AI image generation powered by CellCog.
- Supports creation of single images, editing, style transfer, consistent character sets, and product photography.
- Enables reference-based images, sets of related images, style matching, and composition-based generation.
- Integrates with CellCog mothership skill for setup and API calls.
- Provides detailed documentation on prompt patterns, image specifications (aspect ratios, sizes, styles), and usage tips.
- Includes example prompts for professional use cases and guidance for both simple and complex image generation workflows.
Metadata
Frequently Asked Questions
What is Image Cog?
AI image generation and photo editing powered by CellCog. Text-to-image, image-to-image, consistent characters, product photography, reference-based generati... It is an AI Agent Skill for Claude Code / OpenClaw, with 13992 downloads so far.
How do I install Image Cog?
Run "/install image-cog" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Image Cog free?
Yes, Image Cog is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Image Cog support?
Image Cog is cross-platform and runs anywhere OpenClaw / Claude Code is available (darwin, linux, windows).
Who created Image Cog?
It is built and maintained by CellCog (@nitishgargiitd); the current version is v1.0.12.
More Skills