Token导航 LogoToken导航TokenDH.com
开发需要联网clawhub未标认证来源可访问clear审计通过

vibe-prompt-compiler-portablevivi 提示编译器可移植

Agent Skill

用于辅助提示词、系统指令、Agent 行为约束和工作流模板的整理。它适合让 Agent 规范任务边界、统一输出格式、拆分操作步骤或优化提示词可复用性。使用时需要保留真实业务约束,不要把示例当硬规则;涉及自动执行、外部工具或高风险操作时,应在提示词中明确确认步骤、权限边界和失败处理方式。

总安装

7,102

周安装

293

GitHub Stars

公开资料未说明

下载量

2,321
OpenClaw

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

MIT-0

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:vibe-prompt-compiler-portable(vivi 提示编译器可移植)
来源仓库:https://github.com/qiyuequkanhai/vibe-prompt-compiler-portable
安装命令:
openclaw skills install vibe-prompt-compiler-portable
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install vibe-prompt-compiler-portable

简介

将自然语言请求编译为跨平台高信号提示词。

  • 适合 Cursor、Claude Code 等多工具统一提示管理场景。
  • 输出结构化指令,提升 AI 工具响应一致性与准确性。
  • 安装命令:openclaw skills install vibe-prompt-compiler-portable。
  • 示例不可作为硬规则,需结合实际业务约束调整。

SKILL.md

name
vibe-prompt-compiler-portable
description
Compile rough natural-language coding requests into structured, high-signal prompts for cross-platform AI coding tools such as Cursor, Claude Code, Codex CLI, Gemini CLI, and generic IDE chat assistants. Use when a user describes a vague implementation request like "build a dashboard", "fix this bug", "add CRUD", or "refactor this page" and you should first transform it into a scoped implementation brief before generating code.

Vibe Prompt Compiler Portable

Turn rough implementation requests into portable coding briefs that can be pasted into almost any AI coding tool.

Default Behavior

Use this skill automatically when a user gives a vague coding request. Do not require the user to run scripts first.

Instead:

  1. Classify the request.
  2. Extract the facts already present.
  3. Turn missing details into explicit assumptions.
  4. Compile a clean internal implementation brief.
  5. Use that brief as the source of truth for the coding response.

Only mention scripts when the user wants portability, reusable CLI commands, or a saved handoff for another tool.

Task Types

Classify into one of these:

  • new-project
  • page-ui
  • crud-feature
  • api-backend
  • bugfix
  • refactor
  • ai-feature
  • architecture-review
  • integration
  • automation-workflow
  • deployment
  • general

Use the narrowest obvious type. Only ask a follow-up question when one missing detail blocks useful progress.

Portable Defaults

Unless the user says otherwise:

  • prefer MVP over over-engineering
  • prefer minimal diffs over broad rewrites
  • do not modify unrelated files
  • avoid adding dependencies unless justified
  • make assumptions explicit
  • keep outputs easy to test and verify
  • state non-goals when they reduce drift
  • prefer evolutionary changes over rewrites for existing systems
  • separate current-task execution from future-state ideas

Main Commands

Default: compile mentally and use the structure directly in the answer or coding workflow.

Compile a prompt via CLI when needed:

python3 scripts/compile_prompt.py --request "<user request>"

Create a handoff brief:

python3 scripts/create_handoff.py --request "<user request>" --output handoff

Extract repository-aware rules:

python3 scripts/extract_repo_rules.py --repo-root .

Useful flags:

  • --task auto
  • --stack "Next.js, Supabase, Tailwind"
  • --audience "运营人员"
  • --mode plan-only
  • --mode build-first-slice
  • --mode bugfix
  • --target-tool codex-cli
  • --language-preset chinese-first
  • --ruleset minimal-diff
  • --repo-root .
  • --auto-repo-rules
  • --output json

Target Tool Usage

When the request is broad, architectural, or likely to drift, prefer creating a handoff-style brief over answering from the raw request.

Cursor / VS Code chat

  • run compile_prompt.py or create_handoff.py
  • paste the output into the chat
  • let the tool implement only the current slice

Claude Code / Codex CLI / Gemini CLI

  • use the compiled prompt or handoff as the source of truth
  • ask the coding tool to plan first for broad tasks
  • ask for minimal fixes for bug work
  • prefer --target-tool, --ruleset, and repo-aware flags for stronger execution constraints

Publish Notes

This skill is now suitable for sharing as a portable coding-brief compiler with:

  • structured prompt compilation
  • stronger execution handoffs
  • Chinese-first output
  • tool-specific presets
  • development rulesets
  • repository-aware rule extraction and merging
  • regression coverage across routing, golden outputs, presets, linting, and repo-aware flows

References

  • Read references/auto-mode.md first for the default automatic behavior.
  • Read references/templates.md for prompt skeletons.
  • Read references/routing.md when classification is unclear.
  • Read references/usage.md for suggested workflows in common IDEs and coding tools.
  • Read references/tool-examples.md for concrete examples in Cursor, Claude Code, Codex CLI, and Gemini CLI.
  • Read references/real-examples.md for realistic request-to-brief transformations, especially for architecture, integration, workflow, and vague coding requests.

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

OpenClaw

73.28%
按下载量换算1,701

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills