# Create AI Project Planning Documents

New projects often start with unclear scope, missing decisions, and weak plans. This skill turns a concept into linked planning documents for Claude, Codex, and Claude Code.

## Install

```bash
npx skillstore add byronwilliamscpa/project-planning
```

## Metadata

- Status: approved
- Slug: byronwilliamscpa-project-planning
- Version: 1.0.0
- Author version: 1.0.0
- Skillstore revision: r1
- Version status: valid
- Tree hash: 13c60d60b62b2f5570238d09d9b3c05cc67bcc737056e910dd39e19091b1c0c9
- Author: ByronWilliamsCPA
- GitHub username: ByronWilliamsCPA
- License: MIT
- Repository: https://github.com/ByronWilliamsCPA/fragrance-rater/tree/main/.claude/skills/project-planning
- Ref: a39a91716eadede5f4cdefd78178fed4e837a128
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands, filesystem
- Quality score: 70
- Public page: https://skillstore.pages.dev/skills/byronwilliamscpa-project-planning
- Manifest: https://skillstore.pages.dev/api/skills/byronwilliamscpa-project-planning/manifest

## Capabilities

- Creates project vision, ADR, technical specification, and roadmap documents.
- Guides document order so later plans reference earlier decisions.
- Provides templates for scope, architecture, data models, APIs, risks, and phases.
- Validates required sections, placeholders, document length, and local links.
- Supports optional consensus review through a configured zen MCP tool.
- Synthesizes completed documents into a unified project plan.

## Use Cases

- Plan a New Product Build: Turn a product idea into scope, architecture choices, technical details, and phased delivery steps.
- Prepare Engineering Handoff: Create concise planning documents that help developers start work with fewer missing decisions.
- Resume AI-Assisted Development: Use planning references and prompting patterns to load only the context needed for the next task.

## Prompt Templates

### Create Initial Plans

```
Create planning documents for this project: [describe the project]. Include a project vision, one ADR, a technical specification, and a roadmap.
```

### Fill Placeholder Documents

```
Find planning files with Awaiting Generation status. Replace them with project-specific content for this project: [describe the project].
```

### Review Planning Consistency

```
Review my project vision, ADRs, technical specification, and roadmap for conflicting assumptions. List actionable fixes before writing changes.
```

### Synthesize Delivery Plan

```
Confirm all planning documents are complete, then synthesize them into PROJECT-PLAN.md. Include phase branches, risks, dependencies, and next steps.
```

## Limitations

- Planning quality depends on accurate project context from the user.
- Consensus review requires an optional MCP server and may send content externally.
- The validation script checks document structure, not implementation correctness.
- Shell and git examples should be reviewed before use in a workspace.

## Best Practices

- Provide concrete constraints, target users, and success metrics before generation.
- Review each generated document before using it to guide implementation.
- Remove confidential details before sending documents to external review tools.

## Anti Patterns

- Do not generate plans from a vague idea without clarifying assumptions.
- Do not treat roadmap estimates as commitments without team review.
- Do not run shell or git examples without checking their workspace impact.

## Security Audit

- Audited at: 2026-07-23T12:02:45.245\+00:00
- Summary: Most static findings are false positives caused by Markdown code spans, template placeholders, relative links, and safe local validation examples. One semantic issue remains: optional consensus review can disclose project planning content to an external MCP or model. No evidence found of prompt injection, malware intent, credential access, or malicious command construction.

## Stats

- Views: 408
- Downloads: 11
- Favorites: 0
- Popularity score: 0