skills.rest
Lingtai-AI
Lingtai-AIOFFICIAL
Explore all skills by this author →
GitHub

preset-skill-custom

Documents configuration of the TUI custom OpenAI-compatible provider preset.

692|60|Updated Mar 15, 2026
Search Tags:#openai-compatible#responses API#provider configuration#reasoning effort#websocket transport#tui preset#llm endpoint
One-click install
npx skills add https://github.com/Lingtai-AI/lingtai --skill preset-skill-custom
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: preset-skill-custom
Source: https://github.com/Lingtai-AI/lingtai/tree/main/tui/internal/preset/skills/lingtai-preset-skill/reference/custom
Command: npx skills add https://github.com/Lingtai-AI/lingtai --skill preset-skill-custom

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Configuring a user-supplied OpenAI-compatible LLM endpoint in the LingTai TUI involves uncertain vision capability, transport selection, and reasoning-effort settings that are easy to misconfigure without official-source guidance.

Core Features & Use Cases

  • Custom Preset Reference: Explains the customPreset() template in tui/internal/preset/preset.go, including the LLM_API_KEY credential slot and user-supplied endpoint behavior.
  • Responses Transport Guidance: Documents the wire_api: "responses" path, the http default, and the websocket option that writes responses_transport: "websocket" for Responses WebSocket v2.
  • Reasoning Effort Settings: Details the thinking manifest values (none through xhigh) and the custom-provider default of high.
  • Use Case: When setting up a self-hosted or third-party OpenAI-compatible endpoint, consult this reference to verify vision capability against the provider's official documentation and choose the correct transport and reasoning effort before saving the manifest.

Quick Start

Explain how to configure the custom provider preset in the TUI, including vision capability, Responses transport, and reasoning effort options.

Frequently Asked Questions about preset-skill-custom

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I configure a custom OpenAI-compatible endpoint in the LingTai TUI?▼

Use the custom preset, which leaves the model empty until configured, stores the credential in the LLM_API_KEY slot, and accepts a user-supplied endpoint. Verify all inherited fields in the TUI source and inspect the saved manifest for the actual configuration.

Does the custom provider preset support vision or image inputs?▼

Vision capability is unknown until the actual provider, model, protocol, and endpoint are identified. The vision tool tries the configured OpenAI-compatible endpoint by default; consult the endpoint's official documentation to confirm support.

What is the responses_transport websocket option in the custom preset?▼

For provider custom with api_compat openai and wire_api responses, selecting websocket writes responses_transport: "websocket" and opts into Responses WebSocket v2. The default http option is omitted from the saved manifest.

What happens if the endpoint does not support Responses WebSocket v2?▼

A matching Kernel version reports that the upstream does not support or rejected Responses WebSocket v2 and does not silently fall back to HTTP. Choose http explicitly when the upstream lacks WebSocket support.

What reasoning effort values can I set for a custom provider?▼

On the explicit responses wire path you can set thinking to none, minimal, low, medium, high, or xhigh. Selecting default removes thinking from the manifest and preserves the Kernel's custom-provider default of high.

Related Skills

Explore More Software Engineering Skills→
syeda-hoorain-alisyeda-hoorain-ali

llm-provider-configuration

Configure the OpenAI Agents SDK with a custom base URL and API key for alternative LLM providers.

Community
Intermediate
vercel-labsvercel-labs
246

vercel-plugin-eval

Run live Claude Code eval sessions for the vercel-plugin and generate coverage reports.

Official
Advanced
simstudioaisimstudioai
29.5k

add-model

Adds a new LLM model entry to the Sim providers registry with specs verified against live provider documentation.

Official
Advanced
chiptoe-svgchiptoe-svg

add-codex

Configure OpenAI Codex as a NanoClaw agent provider with flexible authentication.

Community
Intermediate