# Structure AI Development With Specs

Complex AI-assisted projects can lose clarity before implementation starts. This skill guides teams through phased specifications, governance, planning, analysis, and task execution.

## Install

```bash
npx skillstore add joseph obrien/89jobrien-spec-driven-development
```

## Metadata

- Status: approved
- Slug: 89jobrien-spec-driven-development
- Version: 1.0.1
- Author version: 1.0.1
- Skillstore revision: r1
- Version status: valid
- Tree hash: 55e141b58636e43dcf7b4bbb624db5ccae02f6ba5689c2378d576d7c8e1ef859
- Author: Joseph OBrien
- GitHub username: 89jobrien
- License: MIT
- Repository: https://github.com/89jobrien/steve/tree/main/steve/skills/spec-driven-development
- Ref: c4037264bbd363c572662d6154a3ab28f5ca4f53
- 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
- Quality score: 81
- Quality tier: silver
- Public page: https://skillstore.pages.dev/skills/89jobrien-spec-driven-development
- Manifest: https://skillstore.pages.dev/api/skills/89jobrien-spec-driven-development/manifest

## Capabilities

- Defines a seven-phase workflow from project constitution to implementation.
- Guides creation of specifications, plans, research notes, data models, and task lists.
- Uses quality gates for clarifications, analysis, and constitution alignment.
- Explains task dependencies, user story grouping, and optional parallel work markers.
- Describes coordination patterns for Claude, Codex, Claude Code, and other coding agents.

## Use Cases

- Turn Feature Ideas Into Requirements: Create stakeholder-focused specifications with scenarios, acceptance criteria, edge cases, and measurable success criteria.
- Coordinate Parallel Delivery: Break planned work into dependency-aware tasks that multiple developers or AI agents can execute with clear handoffs.
- Prepare AI Coding Sessions: Move from clarified goals to implementation-ready tasks before asking an AI coding assistant to modify files.

## Prompt Templates

### Start a Feature Specification

```
Use spec-driven-development to create a stakeholder-focused specification for [feature]. Ask clarification questions before planning.
```

### Clarify Missing Requirements

```
Review my specification for [feature]. Identify the highest-impact ambiguities and ask one question at a time.
```

### Create an Implementation Plan

```
Use the SDD planning phase for [feature]. Produce the technical context, research needs, design artifacts, and quality gates.
```

### Coordinate Multi-Agent Execution

```
Create a dependency-ordered task breakdown for [feature]. Mark independent tasks and define handoff points for multiple agents.
```

## Limitations

- It provides methodology guidance, not executable automation.
- It does not include the referenced scripts or templates as files.
- Users must adapt artifacts to their project stack and team rules.
- Security, compliance, and architecture decisions still require human review.

## Best Practices

- Establish the constitution before creating feature plans or implementation tasks.
- Resolve high-impact clarification questions before selecting technologies or writing tasks.
- Run analysis before implementation to catch inconsistencies across specs, plans, and tasks.

## Anti Patterns

- Starting implementation before the specification has clear acceptance criteria.
- Creating parallel tasks without explicit dependencies and handoff points.
- Treating the methodology as a substitute for security, compliance, or architecture review.

## Security Audit

- Audited at: 2026-07-05T13:52:07.258\+00:00
- Summary: All static findings were false positives caused by Markdown backticks, example directory trees, and descriptive mentions of Bash or PowerShell script variants. No evidence of executable code, command invocation, reconnaissance behavior, prompt injection, data exfiltration, or malicious intent was found in SKILL.md.

## Stats

- Views: 244
- Downloads: 19
- Favorites: 0
- Popularity score: 0