chat-liquid
GitHub调用 Liquid AI 边缘模型 LFM2.5-1.2B 提供快速、低成本的单轮对话回复。适用于无需复杂推理或代码生成的场景，优先追求速度与成本效益。支持自定义系统指令，若回答质量不足可路由至更强模型。
触发场景
安装
npx skills add zerogpu/zerogpu-router --skill chat-liquid -g -y
SKILL.md
Frontmatter
{
"name": "chat-liquid",
"description": "Fastest, cheapest chat reply via Liquid AI's edge model (LFM2.5-1.2B-Instruct). Use when the user wants a quick, single-turn answer that does not need Claude-level reasoning, prior conversation context, or code generation, and explicitly favors speed or cost over quality. Optional system instructions via -i.",
"allowed-tools": "Bash(zerogpu chat *)",
"argument-hint": "<text> [-i <instructions>]"
}
Call the Liquid AI edge model. $ARGUMENTS is the raw prompt. Pass it verbatim, with no escaping or quoting required (the heredoc below handles every shell metacharacter, newline, quote, and paren safely):
ZGPU_TEXT=$(cat <<'ZGPU_END_OF_INPUT'
$ARGUMENTS
ZGPU_END_OF_INPUT
)
zerogpu chat "$ZGPU_TEXT" -m LFM2.5-1.2B-Instruct
If the user supplied system instructions, append -i "<instructions>" after the model flag.
Output is the assistant's answer as plain text. Relay it as-is, without rewriting or expanding it. At 1.2B parameters this model is terse and occasionally mechanical; that is expected. If the answer is too weak for the task, /zerogpu-router:chat runs the same prompt on gpt-oss-120b.
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.
版本历史
- cee9321 当前 2026-08-02 21:14
