← 返回 Skills 市场
hickhe

Ai Plan Generator

作者 roboe · GitHub ↗ · v2.1.1 · MIT-0
darwinlinuxwin32 ⚠ suspicious
124
总下载
1
收藏
0
当前安装
4
版本数
在 OpenClaw 中安装
/install ai-plan-generator
功能描述
Generates comprehensive campaign documents, task decompositions, and context documents from minimal input for ClawTeam continuous iteration. Supports Code Ar...
使用说明 (SKILL.md)

AI Plan Generator v2

AI Plan Generator v2 extends the original business functionality decomposition with comprehensive campaign document generation, task decomposition, context document generation, and Code Archaeology integration for ClawTeam continuous iteration.

Core Capabilities

1. Campaign Document Generation

  • Minimal Input Processing: Only 4 key inputs required (project name, business goal, scope boundary, code location)
  • Multi-Language Support: Java, Python, Go, C#, Rust, JavaScript
  • Project Type Detection: Automatically identifies new vs iteration projects
  • Domain-Specific Templates: Finance, User Management, General domains

2. Task Decomposition

  • Detailed Task Generation: Complete tasks with priorities, dependencies, acceptance criteria
  • Multiple Output Formats: JSON, Markdown, ClawTeam formats
  • Domain-Specific Tasks: Finance (payment, invoicing, reconciliation), User (auth, permissions)
  • Source File Location: Precise source code location tracking

3. Context Document Generation

  • Business Rules: AI-executable business constraints and validation rules
  • Technical Specifications: Data models, API contracts, integration specifications
  • Validation Standards: Comprehensive testing requirements and coverage criteria
  • Integration Configuration: External system configuration with timeout/retry settings

4. Code Archaeology Integration

  • Unified Directory Structure: results/, process/, source/ subdirectories
  • Real Analysis Results: Extracts precise information from actual Code Archaeology output
  • Completeness Analysis: Validates document completeness and AI executability
  • Clarification Questions: Automatically generates questions for missing/incomplete information

Unified Workflow

Complete End-to-End Process

# 1. Generate complete workflow from minimal input
ai-plan-generator generate-complete-workflow input.json /path/to/code-archaeology

# 2. Creates standardized directory structure:
project-name/
├── project-name-campaign.md          # Campaign document
├── task-decomposition/              # Task decomposition  
│   ├── tasks.json                   # JSON format
│   ├── tasks.md                     # Markdown format
│   └── clawteam-tasks.json          # ClawTeam format
├── context-documents/               # Context documents
│   ├── business-rules.json          # Business rules
│   ├── technical-specs.yaml         # Technical specifications
│   ├── validation-standards.md      # Validation standards
│   ├── integration-config.json       # Integration configuration
│   └── analysis-report.json         # Completeness analysis report
└── process-files-report.json        # Process file location report

Code Archaeology Integration

# Convert Code Archaeology results to AI Plan Generator format
ai-plan-generator generate-context-from-archaeology \
  /path/to/project_code_archaeology \
  context-documents \
  finance

Minimal Input Requirements

Only 4 key pieces of information are required:

{
  "projectName": "dms-erp-finance-migration-v1",
  "businessGoal": "Migrate financial module to Java", 
  "scopeBoundary": "Backend services only, no frontend",
  "codeLocation": "src/main/java/com/dms/financialmanagement/"
}

Domain-Specific Intelligence

Financial Domain

  • Payment Processing: Contract payment handling with amount validation
  • Invoice Generation: Tax calculation, invoice types, compliance
  • Account Reconciliation: Partial refunds, multiple reconciliations
  • Security Requirements: Hardcoded credential removal, SQL injection prevention

User Management Domain

  • Authentication: Registration, login, session management
  • Authorization: Role-based permissions, fine-grained access control
  • Security: Password complexity, OAuth integration, CSRF protection

ClawTeam Integration

Generated artifacts are directly compatible with ClawTeam:

  • Campaign Documents: Used as team description for clawteam create
  • Task Decomposition: Directly importable as clawteam task create commands
  • Context Documents: Provide AI-executable reference for agent implementation
  • Completeness Analysis: Ensures high-quality input before team creation

Best Practices

Input Quality

  • Project Name: Include domain keywords (finance, user, order)
  • Business Goal: Be specific about core business value
  • Scope Boundary: Clearly define inclusions/exclusions
  • Code Location: Use complete package paths

Iteration Strategy

  • Start Simple: Begin with minimal input, let AI infer details
  • Validate Early: Run completeness analysis before team creation
  • Clarify Issues: Address high-priority clarification questions
  • Iterate Refinement: Use feedback to improve subsequent iterations

Example Use Cases

Legacy PHP to Java Migration

Input: Financial module migration from zbs_php to dms-erp Output: Complete campaign document with security remediation, task decomposition with data migration tasks, context documents with precise business rules

New Microservice Development

Input: Create new user authentication service Output: Campaign document with architecture decisions, task decomposition with OAuth integration, context documents with security requirements

API Modernization

Input: Standardize legacy RPC APIs to RESTful design Output: Campaign document with versioning strategy, task decomposition with backward compatibility, context documents with API specifications

AI Plan Generator v2 transforms minimal input into comprehensive, AI-executable documentation that powers successful ClawTeam continuous iteration.

安全使用建议
This skill appears to do what it says (generate campaign, task, and context documents from Code Archaeology output), but it contains hard-coded, user-specific filesystem references and strong defaults for a particular legacy project (zbs_php / dms-erp). Before installing or running it: 1) Inspect the code (especially ai_plan_generator_v2.py and code-archaeology-integrator.js) for any absolute/hidden paths and decide whether they should be removed or made configurable; 2) Run it in a sandboxed environment or with a throwaway user account to observe what local files it reads; 3) Provide only intended archaeology directories on the command line rather than relying on any implicit defaults; 4) If you rely on this for different projects, consider requesting the author generalize the zbs_php-specific filename expectations; 5) If you are uncomfortable with the tool reading your workspace, do not install it or edit the source to remove the hard-coded paths. Additional information that would reduce concern: confirmation that at runtime the tool only reads directories explicitly passed by the user (no fallback to /Users/admin or other user-specific paths) and configurable file-name patterns instead of hard-coded zbs_php_* filenames.
功能分析
Type: OpenClaw Skill Name: ai-plan-generator Version: 2.1.1 The ai-plan-generator skill bundle is a project management and documentation tool designed to transform code analysis results into structured development plans. It consists of Python and Node.js scripts that parse Markdown and JSON files (from a 'Code Archaeology' process) to generate campaign documents, task decompositions, and technical specifications. The code operates entirely within the local OpenClaw workspace (e.g., `/Users/admin/.openclaw/workspace`), lacks external network dependencies, and contains no evidence of malicious intent, data exfiltration, or unauthorized execution. Its primary function is to automate the creation of 'ClawTeam' compatible tasks and business rules for software migration and modernization projects.
能力评估
Purpose & Capability
The skill's name/description (AI plan generation + Code Archaeology integration) is consistent with included JS/Python code and the declared required binaries (node, git, bash). However multiple code files are tightly tailored to a specific Code Archaeology output naming scheme (zbs_php_*) and enterprise conventions (dms-erp, zbs_php). That degree of hard-coding to a particular project/format is narrower than the generic description and could surprise users expecting a broadly reusable tool.
Instruction Scope
SKILL.md directs the agent to run ai-plan-generator CLI with a user-specified archaeology path (expected). But the included Python and JS code also reference absolute/local paths and filenames outside the provided input (not declared in requires/config): e.g., ai_plan_generator_v2.py checks /Users/admin/.openclaw/workspace/zbs_php_security_audit_results.md and the integrator expects specific zbs_php_*.md filenames. These hard-coded reads of a user's workspace are not documented in SKILL.md and expand the skill's data access beyond the minimal input described.
Install Mechanism
No install spec is provided (instruction-only install behavior), and the package just requires common system binaries (node, git, bash). No external downloads or archive extraction are present in the manifest — this is low risk from an install-supply perspective.
Credentials
The skill requests no environment variables or credentials, which is proportional and reduces secret-exfiltration risk. However, code inspects specific local file paths (user workspace) and assumes enterprise integration defaults (e.g., default integration configs and domain assumptions). That implicit access to local files is not declared as required config paths and therefore is disproportionate to the advertised minimal-input model.
Persistence & Privilege
The skill is not always-enabled and does not request system-wide persistence or modification of other skills. Autonomous invocation is allowed (platform default) but not combined with other high-privilege requests.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install ai-plan-generator
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /ai-plan-generator 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v2.1.1
Added comprehensive campaign document generation, task decomposition, context document generation, and Code Archaeology integration with unified directory structure support
v2.1.0
Added comprehensive campaign document generation, task decomposition, context document generation, and Code Archaeology integration with unified directory structure support
v2.0.0
Major update: Focus on business functionality decomposition rather than language-specific implementation. Language-agnostic task generation for flexible AI execution.
v1.0.0
Initial release: Converts Code Archaeology analysis results into AI-executable development task lists with markdown and JSON output formats
元数据
Slug ai-plan-generator
版本 2.1.1
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 4
常见问题

Ai Plan Generator 是什么?

Generates comprehensive campaign documents, task decompositions, and context documents from minimal input for ClawTeam continuous iteration. Supports Code Ar... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 124 次。

如何安装 Ai Plan Generator?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install ai-plan-generator」即可一键安装,无需额外配置。

Ai Plan Generator 是免费的吗?

是的,Ai Plan Generator 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Ai Plan Generator 支持哪些平台?

Ai Plan Generator 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(darwin, linux, win32)。

谁开发了 Ai Plan Generator?

由 roboe(@hickhe)开发并维护,当前版本 v2.1.1。

💬 留言讨论