captchas-openclaw
OpenClaw integration guidance for CAPTCHAS Agent API, including OpenResponses tool schemas and plugin tool registration.
Render JSON schemas to images and generate schemas from prompts using declare-render and AI.
Follow the repository instructions for this skill and your agent. Review the files and dependencies before adding it to your workflow.
SkillHub CLI
skillhub install cai-zhuo/test-materialsInspect the source and compatibility requirements, then follow its installation steps.
Give your agent the source link and ask it to inspect instructions and dependencies before explaining the installation steps.
Use this skill when the user wants to render JSON schemas to images (PNG/JPG), validate render-data schemas, or generate schemas from natural-language prompts and then render them.
Run via Node (from the project or after npm install -g materials-cli):
materials render <schema-path> [options]
materials generate "<prompt>" [options]
materials validate <schema-path> [options]
materials render schema.json -o output.png
-s, --schema <path>, -o, --output <path> (default ./output.png), -f, --format <png|jpg>, -w, --width, -h, --height, --output-schema <path>, -i, --interactive
materials generate "A red circle with text Hello" -o out.png
-o, --output, -f, --format, -w, --width, -h, --height, --output-schema, --model, --api-key, --base-url, -i, --interactive
OPENAI_API_KEY (and optionally OPENAI_MODEL, OPENAI_BASE_URL) if not passed via flags.materials validate schema.json
-s, --schema <path>, -i, --interactive
Usage: materials <command> [options]
Commands:
render <schema> Render a JSON schema file to an image
generate <prompt> Use AI to generate a schema, then render
validate <schema> Validate a schema against the render data schema
Examples:
materials render schema.json -o output.png
materials generate "A red circle with text Hello"
materi...
These signals come from indexed source analysis and may lag behind changes. Check the current source files.
# Materials CLI Use this skill when the user wants to render JSON schemas to images (PNG/JPG), validate render-data schemas, or generate schemas from natural-language prompts and then render them. ## Commands - **render** — Render a JSON schema file to an image. - **generate** — Use AI (OpenAI) to generate a schema from a prompt, then render it. - **validate** — Validate a JSON schema against the declare-render data schema. ## When to use - User asks to "render a schema to image", "turn JSON into a picture", or "draw from schema". - User wants to "generate an image from a description" or "create a schema from a prompt" and render it. - User wants to "validate" a JSON file against the render data schema. ## Usage Run via Node (from the project or after `npm install -g materials-cli`): ```bash materials render <schema-path> [options] materials generate "<prompt>" [options] materials validate <schema-path> [options] ``` ### Render - `materials render schema.json -o output.png` - Options: `-s, --schema <path>`, `-o, --output <path>` (default `./output.png`), `-f, --format <png|jpg>`, `-w, --width`, `-h, --height`, `--output-schema <path>`, `-i, --interactive` ### Generate (AI) - `materials generate "A red circle with text Hello" -o out.png` - Options: `-o, --output`, `-f, --format`, `-w, --width`, `-h, --height`, `--output-schema`, `--model`, `--api-key`, `--base-url`, `-i, --interactive` - Uses `OPENAI_API_KEY` (and optionally `OPENAI_MODEL`, `OPENAI_BASE_URL`) if not passed via flags. ### Validate - `materials validate schema.json` - Options: `-s, --schema <path>`, `-i, --interactive` ## CLI help ``` Usage: materials <command> [options] Commands: render <schema> Render a JSON schema file to an image generate <prompt> Use AI to generate a schema, then render validate <schema> Validate a schema against the render data schema Examples: materials render schema.json -o output.png materials generate "A red circle with text Hello" materi...
OpenClaw integration guidance for CAPTCHAS Agent API, including OpenResponses tool schemas and plugin tool registration.
Register, communicate, and earn on the x402hub AI agent marketplace. Use when an agent needs to register on x402hub, browse or claim bounties, submit deliverables, send messages to other agents via x402 Relay,...
Task Router
name: modelready description: Start using a local or Hugging Face model instantly, directly from chat. metadata: {"openclaw":{"requires":{"bins": "bash", "curl" }, "env": "URL" }}