PoYo Claude Fable 5 Messages
PoYo Links
- Model page: https://poyo.ai/models/claude-fable-5
- API docs: https://docs.poyo.ai/api-manual/chat-series/claude-messages
- API key page: https://poyo.ai/dashboard/api-key
Use this skill for Claude Fable 5 requests through PoYo's Claude-compatible Messages API. It helps agents prepare messages, system instructions, content blocks, tool definitions, structured output settings, streaming calls, and server-side integration notes.
Use When
- The user explicitly asks for PoYo,
claude-fable-5, the PoYo Claude Messages API, or a PoYo/v1/messagesintegration. - The user wants to use that PoYo integration for code assistance, research synthesis, document analysis, knowledge work, multi-turn chat, vision input, or agent tool planning.
- The workflow needs a server-side PoYo curl example, request payload, streaming handler, tool schema, or response parser.
Model Selection
claude-fable-5: use for Claude Fable 5 Messages API requests on PoYo.
Key Inputs
modelis required and should beclaude-fable-5.messagesis required and should containroleandcontent.max_tokenscontrols the maximum response length.systemsets assistant behavior and can use a string or documented content blocks.toolsandtool_choiceconfigure tool use when the application can safely execute tools.output_configcan request structured output when supported.cache_controlcan mark reusable content when supported by the current API.stream: truerequires streaming-aware client handling.
Security Rules
- Treat
POYO_API_KEYas a secret. - Keep PoYo API keys in server-side environment variables or a backend secret manager.
- Never place an API key in browser code, frontend bundles, public repositories, logs, screenshots, or chat output.
- Do not log private messages, system prompts, image content, tool inputs, raw request bodies, or API key headers unless product policy explicitly allows it.
- Do not make live API calls unless the user explicitly asks and provides a safe server-side environment.
- Execute returned tool calls only through an application-controlled allowlist with validated arguments.
Execution
- Read
references/api.mdfor endpoint details, request fields, examples, and response notes. - Use
scripts/submit_claude_fable_5_messages.shonly when the user wants to submit a prepared JSON payload from a trusted shell. - If the user only needs a curl example, adapt one from
references/api.md. - Messages calls are synchronous unless streaming is enabled.
Output Expectations
When helping with Claude Fable 5, include:
- chosen model id
- final payload or concise parameter summary
- synchronous or streaming handling
- system prompt, content blocks, tools, structured output, and cache settings when relevant
- response parsing notes if the user needs integration code
