Use when designing or implementing states for AI agent tasks, background jobs, local model runs, queues, and notifications — one shared set of states, only legal transitions, and a recovery path out of every failure.
中文简介正在准备中。当前显示原文说明。
Use when creating a new project folder, starter pack, app scaffold, or agent-ready workspace — AGENTS.md, README, TASKS, DECISIONS, PROMPTS, an idempotent bootstrap script, and starter-pack modes.
中文简介正在准备中。当前显示原文说明。
实测评分 未实测
查看安装方法安装前可以查看交给 AI 代理的指令内容。
Scaffold a new project so any coding agent can pick it up immediately: structure, memory, tasks, prompts, and clear instructions. Never overwrite user files or copy secrets by default.
Use this skill when creating a new project folder, starter pack, app scaffold, or agent-ready workspace.
Use for project starter apps, bootstrap scripts, universal project templates, and new agent project setup.
Every new project should start with structure, memory, tasks, prompts, and clear instructions for the coding agent.
Use when the user says:
Do not use this skill for:
ProjectName/
AGENTS.md
README.md
TASKS.md
DECISIONS.md
PROMPTS.md
CHANGELOG.md
.env.example
.gitignore
docs/
product-brief.md
architecture.md
ui-notes.md
prompts/
build.md
review.md
polish.md
skills/
src/
assets/
tests/
# AGENTS.md
## Project Goal
Describe what this project is building.
## Rules for the agent
- Read README.md first.
- Check TASKS.md before coding.
- Record major decisions in DECISIONS.md.
- Keep changes small and explain them.
- Run available tests or explain why not.
## Current Stack
- Platform:
- Language:
- Framework:
- Package manager:
## Output Requirements
After each task, report:
1. Files changed
2. What changed
3. How to test
4. Known risks
# Project Name
## Purpose
## Features
## Setup
## Run
## Test
## Notes
# TASKS.md
## Now
- [ ] Define first task
## Next
- [ ] Add next task
## Later
- [ ] Add future idea
## Done
- [ ] Completed work
# DECISIONS.md
## Decision Log
### YYYY-MM-DD — Decision title
Context:
Decision:
Reason:
Tradeoffs:
# PROMPTS.md
## Build Prompt
## Review Prompt
## Polish Prompt
minimal = AGENTS, README, TASKS, DECISIONS, PROMPTS
app = minimal + src, assets, tests, docs
swiftui = app + Xcode notes, app architecture docs
web = app + package scripts and frontend docs
ai = app + prompts, skills, memory, evals
If creating a script:
Input: target folder path
Optional: include .env.local or not
Creates missing folders
Copies template files
Does not overwrite without confirmation
Prints next steps
.env.example, not .env.local.#!/usr/bin/env bash
set -euo pipefail
target="${1:-}"
if [[ -z "$target" ]]; then
echo "Usage: ./bootstrap.sh /path/to/new-project"
exit 1
fi
mkdir -p "$target"/{docs,prompts,skills,src,assets,tests}
touch "$target/README.md"
touch "$target/TASKS.md"
touch "$target/DECISIONS.md"
touch "$target/PROMPTS.md"
touch "$target/.env.example"
echo "Created project starter at: $target"
For a menu bar starter app:
Settings:
- Default projects folder
- Starter files to include
- Skill folders to include
- Default AGENTS.md template
- Git init toggle
- Open in Finder toggle
- Open in editor toggle
Create flow:
- Project name
- Location
- Starter pack mode
- Selected skills
- Create
[ ] Project has AGENTS.md
[ ] Project has README.md
[ ] Project has TASKS.md
[ ] Project has DECISIONS.md
[ ] Project has PROMPTS.md
[ ] Secrets are not copied by default
[ ] Existing files are not overwritten silently
[ ] Folder structure matches project type
[ ] Next steps are printed
[ ] The agent can understand what to do next
Starting projects with only source files
No AGENTS.md
No task list
No decision log
Copying .env.local by default
Overwriting files silently
Too many folders with no purpose
No README setup instructions
Drop this into any coding agent (Claude Code, Codex, Cursor, …) to apply the skill:
Use the agent-starter-pack skill to create an agent-ready project starter.
Rules:
- Create a clear project folder structure.
- Include AGENTS.md, README.md, TASKS.md, DECISIONS.md, PROMPTS.md, and .env.example.
- Add docs/, prompts/, skills/, src/, assets/, and tests/ as appropriate.
- Do not overwrite existing files silently.
- Do not copy secrets by default.
- Make bootstrap scripts idempotent.
- Print created and skipped files.
- Add clear next steps for the user.
After coding:
1. List created files/folders.
2. Explain starter pack mode.
3. Explain overwrite/secrets behavior.
4. Give manual test steps.
还没有评价。欢迎分享你的使用感受。
无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。
还没有评论。来写第一条吧。
Use when designing or implementing states for AI agent tasks, background jobs, local model runs, queues, and notifications — one shared set of states, only legal transitions, and a recovery path out of every failure.
中文简介正在准备中。当前显示原文说明。
Use when exposing app features to App Intents, Shortcuts, Siri, Spotlight, widgets, controls, or system actions — small safe AppIntents, AppShortcutsProvider, AppEnum/AppEntity, parameters, and open-vs-background behavior.
中文简介正在准备中。当前显示原文说明。
Use when designing, reviewing, or implementing a command-palette / ⌘K layer for an Apple-platform app — keyboard-first global search, ranked and grouped results, native floating panel, and protected destructive commands.
中文简介正在准备中。当前显示原文说明。
Use when designing or implementing empty, first-run, no-result, offline, error, or waiting states in Apple-platform apps — every empty state explains what is missing, why it matters, and what to do next.
中文简介正在准备中。当前显示原文说明。
Use when designing or implementing feedback, status, progress, alerts, errors, success states, and live activity indicators in Apple-platform apps — inline status, banners, alerts, notifications, and loading states.
中文简介正在准备中。当前显示原文说明。
Use when designing, reviewing, or implementing an Inspector panel for an Apple-platform app — contextual selected-item details and safe editable properties, grouped sections, ⌘I toggle, and a compact-width fallback.
中文简介正在准备中。当前显示原文说明。