chat-thinking
GitHub调用 ZeroGPU LFM2.5-1.2B-Thinking 模型，对逻辑或数学问题提供显式推理步骤。适用于需要展示思考过程的用户请求。
Trigger Scenarios
Install
npx skills add zerogpu/zerogpu-router --skill chat-thinking -g -y
SKILL.md
Frontmatter
{
"name": "chat-thinking",
"description": "Chat with ZeroGPU's Thinking variant (LFM2.5-1.2B-Thinking), which returns visible reasoning alongside the answer. Use when the user wants the model's reasoning shown, or asks a short logic\/math\/word-problem question that benefits from step-by-step output.",
"allowed-tools": "Bash(zerogpu chat_completions *)",
"argument-hint": "<text>"
}
Call the ZeroGPU thinking model. $ARGUMENTS is the raw user prompt — pass it verbatim, no escaping or quoting required (the heredoc below handles every shell metacharacter, newline, quote, and paren safely):
zerogpu chat_completions -m LFM2.5-1.2B-Thinking --raw <<'ZGPU_END_OF_INPUT'
$ARGUMENTS
ZGPU_END_OF_INPUT
Output is the full Chat Completions response as JSON, because the reasoning is a separate field there. Show the model's reasoning first — choices[0].message.reasoning, or reasoning_content if that is the field present — then its answer from choices[0].message.content. If there is no separate reasoning field, the content carries the reasoning inline; show it as-is. Do not print the rest of the JSON.
Savings note: only if the command output literally contains a line starting with 💰 ZeroGPU savings, append that exact line, unchanged, as the last line of your reply. If no such line is present, say nothing about savings and do not mention or suggest /zerogpu-router:cost-savings — this note is intentionally occasional, not shown every time.
Version History
-
4e1b070
Current 2026-09-22 08:06
v2.3.0：从专用命令迁移至 zerogpu-cli 3.8.0 的通用 chat_completions 端点，提升插件独立性。
- 7b9b503 2026-07-24 11:47
