AgentSkillsCN

aimlapi-llm-reasoning

运行AIMLAPI（兼容OpenAI）大语言模型与推理工作流，包括聊天补全、工具式提示词，以及结构化输出。当Codex需要编写提示词、调用大语言模型，或通过AIMLAPI端点实现推理与分析的脚本化时，便可使用该功能。

SKILL.md
--- frontmatter
name: aimlapi-llm-reasoning
description: Run AIMLAPI (OpenAI-compatible) LLM and reasoning workflows, including chat completions, tool-style prompts, and structured outputs. Use when Codex needs to craft prompts, run LLM calls, or script reasoning/analysis via the AIMLAPI endpoint.

AIMLAPI LLM + Reasoning

Overview

Use reusable scripts to call AIMLAPI chat completions, pass model-specific reasoning parameters, and capture structured outputs.

Quick start

bash
export AIMLAPI_API_KEY="sk-aimlapi-..."
python3 {baseDir}/scripts/run_chat.py --model aimlapi/openai/gpt-5-nano-2025-08-07 --user "Summarize this in 3 bullets."

Tasks

Run a basic chat completion

bash
python3 {baseDir}/scripts/run_chat.py \
  --model aimlapi/openai/gpt-5-nano-2025-08-07 \
  --system "You are a concise assistant." \
  --user "Draft a project kickoff checklist."

Add reasoning or provider-specific parameters

Use --extra-json to pass fields such as reasoning effort, response format, or tool configs without editing the script.

bash
python3 {baseDir}/scripts/run_chat.py \
  --model aimlapi/openai/gpt-5-nano-2025-08-07 \
  --user "Plan a 5-step rollout for a new chatbot feature." \
  --extra-json '{"reasoning": {"effort": "medium"}, "temperature": 0.3}'

Structured JSON output

bash
python3 {baseDir}/scripts/run_chat.py \
  --model aimlapi/openai/gpt-5-nano-2025-08-07 \
  --user "Return a JSON array of 3 project risks with mitigation." \
  --extra-json '{"response_format": {"type": "json_object"}}'

References

  • •references/aimlapi-llm.md: payload fields and troubleshooting notes.