Token导航 LogoToken导航TokenDH.com
AI 工具需要联网github未标认证来源可访问clear审计通过

project-init项目初始化

Agent Skill

project-init 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

1,270

周安装

54

GitHub Stars

14

下载量

445
CodexClaudeCursorGemini CLI

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

3

许可证

MIT

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:project-init(项目初始化)
来源仓库:https://github.com/vladm3105/aidoc-flow-framework
仓库路径:skills/project-init
安装命令:
npx skills add https://github.com/vladm3105/aidoc-flow-framework --skill project-init
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。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/vladm3105/aidoc-flow-framework --skill project-init

简介

project-init 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中围绕仓库状态、代码变更或协作事项进行整理。

  • 适用于项目初始化、环境搭建和协作流程启动等 AI 工具类任务。
  • 可结合来源仓库和安装命令进一步核验具体用法和功能边界。
  • 安装前建议确认权限范围和维护状态,避免触发不必要的联网或命令执行。
  • 需注意是否会读写文件或访问私有仓库,确保 token 权限和用户授权合规。

SKILL.md

project-init

Purpose

Provide AI assistants with structured guidance for initializing brand new (greenfield) projects using the AI Dev Flow framework. This skill handles the one-time setup process that must occur BEFORE workflow execution begins.

⚠️ CRITICAL: This skill is for NEW PROJECT INITIALIZATION ONLY. For ongoing workflow execution on existing projects, use the doc-flow skill instead.

When to Use This Skill

Use project-init when:

  • Starting a brand new project from scratch (greenfield)
  • No project folders exist yet
  • Domain has not been selected
  • Project structure needs to be initialized

Do NOT use project-init when:

  • Project already has docs/ folder structure
  • Domain is already configured
  • Working on existing project
  • → Use doc-flow skill instead for workflow execution

Hand-off to doc-flow

After completing project initialization, AI Assistant MUST inform user:

✅ Project initialization complete!

Next: Use the `doc-flow` skill to begin workflow execution:
- Create BRD (Business Requirements)
- Create PRD (Product Requirements)
- Follow 15-layer architecture (Layers 0-14) with 11 artifact directories (BRD through TASKS): BRD → PRD → EARS → BDD → ADR → SYS → REQ → [IMPL] → [CTR] → SPEC → TASKS → Code → Tests → Validation

Initialization Workflow (8 Steps)

Step 0: Read Execution Rules

File: AI_ASSISTANT_RULES.md

Purpose: Understand core execution rules before starting

Key Rules:

  1. Domain Selection FIRST
  2. Create Folders BEFORE Documents
  3. Apply Domain Configuration
  4. Run Contract Questionnaire
  5. Initialize Index Files
  6. Validate Setup
  7. Token Optimization
  8. Tool-Specific Guidance

Action: AI Assistant reads this file to understand all 15 rules


Step 1: Domain Selection (REQUIRED FIRST)

File: DOMAIN_SELECTION_QUESTIONNAIRE.md

Purpose: Determine project domain to load correct configuration

AI Assistant Action: Present questionnaire to user

═══════════════════════════════════════════════════════════
                 PROJECT DOMAIN SELECTION
═══════════════════════════════════════════════════════════

What is the purpose and focus of this new project?

Select one:

1. Financial Services (DEFAULT)
   - Trading platforms, banking, insurance, portfolio management
   - Regulatory: SEC, FINRA, SOX, Basel III, PCI-DSS

2. Software/SaaS
   - B2B/B2C software services, multi-tenant applications
   - Regulatory: SOC2, GDPR/CCPA, ISO 27001

3. Healthcare
   - EMR, telemedicine, medical devices, patient management
   - Regulatory: HIPAA, FDA, HITECH, 21 CFR Part 11

4. E-commerce
   - Retail, marketplace, subscription services
   - Regulatory: PCI-DSS, GDPR/CCPA, FTC

5. IoT (Internet of Things)
   - Connected devices, sensors, industrial systems
   - Regulatory: FCC, CE, UL/IEC, FDA (medical devices)

6. Other/Generic
   - Internal tools, utilities, custom domain
   - Regulatory: Company policies only

Enter selection (1-6) or press Enter for default (1):

Output: Domain selected, configuration file determined

Domain Configuration Mapping:

SelectionDomainConfig File
1 or EnterFinancial ServicesFINANCIAL_DOMAIN_CONFIG.md
2Software/SaaSSOFTWARE_DOMAIN_CONFIG.md
3HealthcareDOMAIN_ADAPTATION_GUIDE.md (Healthcare)
4E-commerceDOMAIN_ADAPTATION_GUIDE.md (E-commerce)
5IoTDOMAIN_ADAPTATION_GUIDE.md (IoT)
6Other/GenericGENERIC_DOMAIN_CONFIG.md

Step 2: Folder Structure Creation (REQUIRED SECOND)

Rule: AI Assistant MUST create complete directory structure BEFORE creating any documents.

Why: Prevents "file not found" errors, ensures proper organization

IMPORTANT: Ensure project root directory exists first:

# Create project root directory if it doesn't exist
mkdir -p {project_root}
cd {project_root}
Replace {project_root} with your actual project path (e.g., /opt/data/my_project)

Commands to Execute:

# Core 11 artifact directories (BRD through TASKS)
mkdir -p docs/BRD
mkdir -p docs/PRD
mkdir -p docs/EARS
mkdir -p docs/BDD
mkdir -p docs/ADR
mkdir -p docs/SYS
mkdir -p docs/REQ
mkdir -p docs/IMPL
mkdir -p docs/CTR
mkdir -p docs/SPEC
mkdir -p docs/TASKS

# NOTE: REQ and CTR subdirectories are created on-demand by doc-req and doc-ctr skills
# when documents are generated. This ensures folders match actual project needs.

# Scripts directory
mkdir -p scripts

# Work plans directory (for /save-plan command output)
mkdir -p work_plans

Validation:

ls -la docs/  # Verify 11 artifact directories created
ls -la work_plans/  # Verify work_plans directory

Step 3: Load Domain Configuration

Files:

Purpose: Apply domain-specific terminology and placeholders

AI Assistant Action:

  1. Read selected domain configuration file
  2. Extract placeholder mappings
  3. Store terminology for document generation

Example Mappings:

Financial Services:

[RESOURCE_COLLECTION] → Portfolio
[RESOURCE_ITEM] → Position
[USER_ROLE] → Trader / Portfolio Manager
[TRANSACTION] → Trade
[REGULATORY_REQUIREMENT] → SEC Rule 15c3-5

Software/SaaS:

[RESOURCE_COLLECTION] → Workspace
[RESOURCE_ITEM] → Resource
[USER_ROLE] → Account Admin / Member
[TRANSACTION] → API Call
[REGULATORY_REQUIREMENT] → SOC2 Control

Generic:

[RESOURCE_COLLECTION] → Collection
[RESOURCE_ITEM] → Entity
[USER_ROLE] → User
[TRANSACTION] → Action
[REGULATORY_REQUIREMENT] → Company Policy

Step 4: Template Copying (Optional)

Purpose: Copy framework templates to project (optional step)

Commands:

# Create ai_dev_flow directory for framework templates
mkdir -p ai_dev_flow

# Copy all templates (if framework templates exist)
cp -r {framework_root}/ai_dev_flow/* ai_dev_flow/

# Copy validation scripts
cp {framework_root}/ai_dev_flow/scripts/*.py scripts/

Directory Purpose:

  • ai_dev_flow/ = Framework templates (BRD-TEMPLATE.md, examples/, etc.)
  • docs/ = Project documentation (BRD-01.md, PRD-01.md, etc.)

Note: This step is optional. Templates can also be referenced directly from framework location.


Step 5: Contract Decision (REQUIRED)

File: CONTRACT_DECISION_QUESTIONNAIRE.md

Purpose: Determine if CTR (Contracts) layer should be included in workflow

AI Assistant Action: Present questionnaire to user

═══════════════════════════════════════════════════════════
              CONTRACT DECISION QUESTIONNAIRE
═══════════════════════════════════════════════════════════

Does this project require API contracts or interface definitions?

Select all that apply:

1. ☐ REST/GraphQL APIs (External HTTP endpoints)
2. ☐ Event Schemas (Pub/Sub, message queues, webhooks)
3. ☐ Data Contracts (Shared database schemas, data models between services)
4. ☐ RPC/gRPC Interfaces (Service-to-service communication)
5. ☐ WebSocket APIs (Real-time bidirectional communication)
6. ☐ File Format Specifications (CSV, JSON, XML exchange formats)
7. ☐ None - Internal logic only
8. ☐ Unsure - Need guidance

Enter selections (comma-separated, e.g., "1,2" or single "7"):

Decision Matrix:

SelectionInclude CTR?Workflow
1-6YESREQ → IMPL → CTR → SPEC → TASKS
7NOREQ → IMPL → SPEC → TASKS
8Ask follow-up questionsSee CONTRACT_DECISION_QUESTIONNAIRE.md

Output: Workflow determined (with or without CTR layer)


Step 6: Index File Initialization + Document Control

Purpose: Create index files for each document type

Document Control Requirements: All AI Dev Flow templates include a Document Control section with:

  • Project metadata (name, version, date, owner, preparer, status)
  • Document Revision History table
  • Essential for traceability, change management, and regulatory compliance

AI Assistant Must Emphasize: When creating documents from templates, users must complete the Document Control section with all required fields.

Commands:

# Create index files
touch docs/BRD/BRD-00_index.md
touch docs/prd/PRD-00_index.md
touch docs/ears/EARS-00_index.md
touch docs/BDD/BDD-00_index.md
touch docs/adrs/ADR-00_index.md
touch docs/sys/SYS-00_index.md
touch docs/REQ/REQ-00_index.md
touch docs/IMPL/IMPL-00_index.md
touch docs/CTR/CTR-00_index.md
touch docs/specs/SPEC-00_index.yaml
touch docs/TASKS/TASKS-00_index.md

Index File Purpose:

  • Track all documents of each type
  • Provide next available ID
  • Document registry with status

Step 7: Validation

Purpose: Verify setup complete and correct

Validation Commands:

# Verify directory structure
ls -laR docs/

# Verify index files exist
ls docs/*/index.* || ls docs/*/*_index.*

# Expected: 11 artifact directories (BRD through TASKS) + domain subdirectories
# Expected: 11 index files

Success Criteria:

  • ✅ All 11 artifact directories exist (BRD, PRD, EARS, BDD, ADR, SYS, REQ, IMPL, CTR, SPEC, TASKS)
  • ✅ Domain-specific subdirectories exist (risk/, trading/, tenant/, etc.)
  • ✅ All index files created
  • ✅ All templates include Document Control sections
  • ✅ Validation scripts present (if copied)
  • ✅ work_plans directory exists (for /save-plan command)

Error Handling:

  • If folders missing: Re-run Step 2
  • If index files missing: Re-run Step 6
  • If domain subdirs missing: Check Step 1 domain selection

Step 8: Project Ready - Hand-off to doc-flow

AI Assistant Confirmation Message:

═══════════════════════════════════════════════════════════
              PROJECT INITIALIZATION COMPLETE
═══════════════════════════════════════════════════════════

✓ Domain: [Financial Services / Software/SaaS / etc.]
✓ Folders: Created (11 artifact directories + domain subdirectories)
✓ Domain Config: Applied ([PLACEHOLDERS] → [domain terms])
✓ Contracts: [Included / Skipped] (CTR layer [active / inactive])
✓ Index Files: Initialized (11 files)
✓ Validation: Passed

Workflow Configuration:
[With CTR]:    REQ → IMPL → CTR → SPEC → TASKS → Code
[Without CTR]: REQ → IMPL → SPEC → TASKS → Code

═══════════════════════════════════════════════════════════
                      NEXT STEPS
═══════════════════════════════════════════════════════════

✅ Project structure ready for development!

Next: Use the `doc-flow` skill to begin workflow execution

Week 1 Tasks (see PROJECT_KICKOFF_TASKS.md):
- Day 1: Create BRD (Business Requirements)
- Day 2: Create PRD + EARS (Product Requirements)
- Day 3: Create BDD + ADR (Tests + Architecture)
- Day 4: Create SYS + REQ (System Specs + Requirements)
- Day 5: Create IMPL + CTR (Implementation Plan + Contracts)
- Day 6: Create SPEC (Technical Specifications)
- Day 7: Create TASKS + Validation

Invoke: doc-flow skill to start Day 1

═══════════════════════════════════════════════════════════

Reference Files

All guidance files located in: {project_root}/ai_dev_flow/

Core Guidance Files

  1. AI_ASSISTANT_RULES.md - 15 execution rules
  2. DOMAIN_SELECTION_QUESTIONNAIRE.md - Domain selection
  3. CONTRACT_DECISION_QUESTIONNAIRE.md - Contract decision
  4. PROJECT_SETUP_GUIDE.md - Master setup guide
  5. PROJECT_KICKOFF_TASKS.md - Week 1 tasks
  6. TRACEABILITY_SETUP.md - Validation automation
  7. QUICK_REFERENCE.md - Quick reference card

Domain Configuration Files

  1. FINANCIAL_DOMAIN_CONFIG.md - Financial Services (DEFAULT)
  2. SOFTWARE_DOMAIN_CONFIG.md - Software/SaaS
  3. GENERIC_DOMAIN_CONFIG.md - Generic/Other

Example Usage

Example 1: Software Application Project

User: "Initialize new software application project"

AI Assistant (using project-init skill):

Step 1: Domain Selection
"What is the purpose and focus of this new project?"

User: "2" (Software/SaaS)

AI Assistant:
✓ Domain selected: Software/SaaS
✓ Loading configuration: SOFTWARE_DOMAIN_CONFIG.md
✓ Domain-specific subdirectories: tenant, subscription, billing, workspace

Step 2: Folder Creation
Creating 15-layer architecture (12 artifact directories)...
✓ Created: docs/BRD/
✓ Created: docs/PRD/
✓ Created: docs/EARS/
✓ Created: docs/BDD/
✓ Created: docs/ADR/
✓ Created: docs/SYS/
✓ Created: docs/REQ/
✓ Created: docs/IMPL/
✓ Created: docs/CTR/
✓ Created: docs/SPEC/
✓ Created: docs/TASKS/
✓ Created: docs/REQ/tenant/
✓ Created: docs/REQ/subscription/
✓ Created: docs/REQ/billing/
✓ Created: docs/REQ/workspace/
✓ Created: scripts/
✓ Created: work_plans/

Step 3: Domain Configuration
Applying Software/SaaS terminology:
✓ [RESOURCE_COLLECTION] → Workspace
✓ [RESOURCE_ITEM] → Resource
✓ [USER_ROLE] → Account Admin
✓ [REGULATORY_REQUIREMENT] → SOC2 Control

Step 5: Contract Decision
"Does this project require API contracts?"

User: "1,2" (REST APIs + Event Schemas)

AI Assistant:
✓ Contracts needed
✓ Contract types: REST API (OpenAPI), Event Schemas (AsyncAPI)
✓ Workflow: REQ → IMPL → CTR → SPEC → TASKS

Step 6: Index Files
✓ All 11 index files created

Step 7: Validation
✓ All directories verified
✓ All index files verified

═══════════════════════════════════════════════════════════
              PROJECT INITIALIZATION COMPLETE
═══════════════════════════════════════════════════════════

✓ Domain: Software/SaaS
✓ Folders: Created (11 artifact directories + 4 SaaS subdirectories)
✓ Contracts: Included (CTR layer active)
✓ Workflow: REQ → IMPL → CTR → SPEC → TASKS

Next: Use `doc-flow` skill to begin workflow execution
Start with: BRD-01 (Business Requirements Document)

═══════════════════════════════════════════════════════════

Example 2: Software/SaaS Multi-Tenant Platform

User: "Set up new SaaS project for team collaboration platform"

AI Assistant (using project-init skill):

Step 1: Domain Selection
User: "2" (Software/SaaS)

AI Assistant:
✓ Domain selected: Software/SaaS
✓ Loading configuration: SOFTWARE_DOMAIN_CONFIG.md
✓ Domain-specific subdirectories: tenant, subscription, billing, workspace

Step 2: Folder Creation
✓ All core directories created
✓ SaaS subdirectories created: tenant/, subscription/, billing/, workspace/
✓ Support directories created: scripts/, work_plans/

Step 3: Domain Configuration
✓ [RESOURCE_COLLECTION] → Workspace
✓ [RESOURCE_ITEM] → Resource
✓ [USER_ROLE] → Account Admin
✓ [TRANSACTION] → API Call
✓ [REGULATORY_REQUIREMENT] → SOC2 Control

Step 5: Contract Decision
User: "1,2" (REST API + Event Schemas)

AI Assistant:
✓ Contracts included
✓ Workflow: REQ → IMPL → CTR → SPEC → TASKS

Project Ready!
Next: Use `doc-flow` skill for BRD-01 creation

Troubleshooting

Issue: Directory Already Exists

Error: mkdir: cannot create directory 'docs': File exists

Resolution:

  • Project may already be initialized
  • Verify: ls -la docs/
  • If folders exist, skip to doc-flow skill
  • If folders incomplete, delete and re-run project-init

Issue: Permission Denied

Error: mkdir: cannot create directory: Permission denied

Resolution:

  • Check current working directory permissions
  • Ensure write access to parent directory
  • Use pwd to verify location

Issue: Domain Unclear

User says: "I'm not sure which domain to choose"

AI Assistant Action: Run follow-up questions from DOMAIN_SELECTION_QUESTIONNAIRE.md:

  1. Does project involve financial transactions? → Financial Services
  2. Multi-tenant SaaS application? → Software/SaaS
  3. Handle patient health information? → Healthcare
  4. Online store or marketplace? → E-commerce
  5. Connected devices or sensors? → IoT
  6. None of above? → Generic

Tool Optimization Notes

Claude Code

  • File limit: 50K tokens (200KB) standard, 100K max
  • Strategy: Single comprehensive files
  • No artificial splitting needed

Gemini CLI

  • @ reference limit: 10K tokens (40KB)
  • Large files: Use file read tool instead of @
  • Command: gemini read FILE.md

GitHub Copilot

  • Optimal: 10-30KB per file
  • Large files: Create companion summaries
  • Working set: Max 10 files in Edits mode

Related Skills

After project-init completes, use:

  • doc-flow - Main workflow execution skill

- Create BRD, PRD, EARS, BDD, ADR, SYS, REQ, IMPL, CTR, SPEC, TASKS - Follow 15-layer architecture (Layers 0-14: Strategy layer + 11 artifact directories + 3 execution layers) - Generate code from specifications

Other complementary skills:

  • mermaid-gen - Generate Mermaid diagrams
  • charts-flow - Create architecture diagrams
  • test-automation - Test suite creation
  • code-review - Code quality review
  • security-audit - Security analysis

Quick Reference

When to use project-init:

  • ✅ Brand new project (no folders exist)
  • ✅ Greenfield development
  • ✅ Starting from scratch

When to use doc-flow:

  • ✅ Project already initialized
  • ✅ Folders exist (docs/BRD/, docs/PRD/, etc.)
  • ✅ Workflow execution (creating BRD, PRD, SPEC, etc.)

Workflow sequence:

project-init (Day 0) → doc-flow (Day 1+) → other skills (as needed)

End of project-init Skill

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

04

需要参考平台分布和安装热度时

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

Gemini CLI

28.44%
按下载量换算127

Claude Code

24.06%
按下载量换算107

OpenCode

16.82%
按下载量换算75

Antigravity

13.02%
按下载量换算58

Codex

8.84%
按下载量换算39

windsurf

3.39%
按下载量换算15

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

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。

来源信息

继续浏览同类 Skills